output
bool 2
classes | input
stringlengths 345
2.91k
|
|---|---|
false |
Differentiable Approximations for Multi-resource Spatial Coverage Problems
Multi-agent coverage Multi-resource coverage Areal coverage Differentiable approximations
Resource allocation for coverage of physical spaces is a challenging problem in robotic surveillance, mobile sensor networks and security domains. Recent gradient-based optimization approaches to this problem estimate utilities of actions by using neural networks to learn a differentiable approximation to spatial coverage objectives. In this work, we empirically show that spatial coverage objectives with multiple-resources are combinatorially hard to approximate for neural networks and lead to sub-optimal policies. As our major contribution, we propose a tractable framework to approximate a general class of spatial coverage objectives and their gradients using a combination of Newton-Leibniz theorem, spatial discretization and implicit boundary differentiation. We empirically demonstrate the efficacy of our proposed framework on single and multi-agent spatial coverage problems.
|
true |
Gradient Origin Networks
Deep Learning Generative Models Implicit Representation
This paper proposes a new type of generative model that is able to quickly learn a latent representation without an encoder. This is achieved using empirical Bayes to calculate the expectation of the posterior, which is implemented by initialising a latent vector with zeros, then using the gradient of the log-likelihood of the data with respect to this zero vector as new latent points. The approach has similar characteristics to autoencoders, but with a simpler architecture, and is demonstrated in a variational autoencoder equivalent that permits sampling. This also allows implicit representation networks to learn a space of implicit functions without requiring a hypernetwork, retaining their representation advantages across datasets. The experiments show that the proposed method converges faster, with significantly lower reconstruction error than autoencoders, while requiring half the parameters.
|
false |
Impact-driven Exploration with Contrastive Unsupervised Representations
reinforcement learning exploration curiosity episodic memory
Procedurally-generated sparse reward environments pose significant challenges for many RL algorithms. The recently proposed impact-driven exploration method (RIDE) by Raileanu & Rocktäschel (2020), which rewards actions that lead to large changes (measured by $\ell_2$-distance) in the observation embedding, achieves state-of-the-art performance on such procedurally-generated MiniGrid tasks. Yet, the definition of "impact" in RIDE is not conceptually clear because its learned embedding space is not inherently equipped with any similarity measure, let alone $\ell_2$-distance. We resolve this issue in RIDE via contrastive learning. That is, we train the embedding with respect to cosine similarity, where we define two observations to be similar if the agent can reach one observation from the other within a few steps, and define impact in terms of this similarity measure. Experimental results show that our method performs similarly to RIDE on the MiniGrid benchmarks while learning a conceptually clear embedding space equipped with the cosine similarity measure. Our modification of RIDE also provides a new perspective which connects RIDE and episodic curiosity (Savinov et al., 2019), a different exploration method which rewards the agent for visiting states that are unfamiliar to the agent's episodic memory. By incorporating episodic memory into our method, we outperform RIDE on the MiniGrid benchmarks.
|
false |
World Model as a Graph: Learning Latent Landmarks for Planning
Reinforcement Learning Planning
Planning, the ability to analyze the structure of a problem in the large and decompose it into interrelated subproblems, is a hallmark of human intelligence. While deep reinforcement learning (RL) has shown great promise for solving relatively straightforward control tasks, it remains an open problem how to best incorporate planning into existing deep RL paradigms to handle increasingly complex environments. One prominent framework, Model-Based RL, learns a world model and plans using step-by-step virtual rollouts. This type of world model quickly diverges from reality when the planning horizon increases, thus struggling at long-horizon planning. How can we learn world models that endow agents with the ability to do temporally extended reasoning? In this work, we propose to learn graph-structured world models composed of sparse, multi-step transitions. We devise a novel algorithm to learn latent landmarks that are scattered (in terms of reachability) across the goal space as the nodes on the graph. In this same graph, the edges are the reachability estimates distilled from Q-functions. On a variety of high-dimensional continuous control tasks ranging from robotic manipulation to navigation, we demonstrate that our method, named L^{3}P, significantly outperforms prior work, and is oftentimes the only method capable of leveraging both the robustness of model-free RL and generalization of graph-search algorithms. We believe our work is an important step towards scalable planning in reinforcement learning.
|
false |
Adversarial Meta-Learning
Meta-learning Adversarial manner Adversarial Meta-Learner Adversarial samples
Meta-learning enables a model to learn from very limited data to undertake a new task. In this paper, we study the general meta-learning with adversarial samples. We present a meta-learning algorithm, ADML (ADversarial Meta-Learner), which leverages clean and adversarial samples to optimize the initialization of a learning model in an adversarial manner. ADML leads to the following desirable properties: 1) it turns out to be very effective even in the cases with only clean samples; 2) it is robust to adversarial samples, i.e., unlike other meta-learning algorithms, it only leads to a minor performance degradation when there are adversarial samples; 3) it sheds light on tackling the cases with limited and even contaminated samples. It has been shown by extensive experimental results that ADML outperforms several representative meta-learning algorithms in the cases involving adversarial samples generated by different attack mechanisms, on two widely-used image datasets, MiniImageNet and CIFAR100, in terms of both accuracy and robustness.
|
false |
Thinking Like Transformers
transformers interpretability
What is the computational model behind a transformer? Where recurrent neural networks have direct parallels in finite state machines, allowing clear discussion and thought around architecture variants or trained models, transformers have no such familiar parallel. In this paper we aim to change that, proposing a computational model for the transformer-encoder in the form of a programming language. We map the basic components of a transformer-encoder – attention and feed-forward computation – into the simple primitives of select, aggregate, and zipmap, around which we form a programming language: the Restricted Access Sequence Process-ing Language (RASP). We show how RASP can be used to program solutions to tasks that could conceivably be learned by a transformer, augmenting it with tools we discover in our work. In particular, we provide RASP programs for histograms, sorting, and even logical inference similar to that of Clark et al. (2020). We further use our model to relate their difficulty in terms of the number of required layers and attention heads. Finally, we see how insights gained from our abstraction might be used to explain phenomena seen in recent works.
|
true |
Decentralized Attribution of Generative Models
GANs Generative Model Deepfake Model Attribution
Growing applications of generative models have led to new threats such as malicious personation and digital copyright infringement.
One solution to these threats is model attribution, i.e., the identification of user-end models where the contents under question are generated.
Existing studies showed empirical feasibility of attribution through a centralized classifier trained on all existing user-end models.
However, this approach is not scalable in a reality where the number of models ever grows. Neither does it provide an attributability guarantee.
To this end, this paper studies decentralized attribution, which relies on binary classifiers associated with each user-end model.
Each binary classifier is parameterized by a user-specific key and distinguishes its associated model distribution from the authentic data distribution.
We develop sufficient conditions of the keys that guarantee an attributability lower bound.
Our method is validated on MNIST, CelebA, and FFHQ datasets. We also examine the trade-off between generation quality and robustness of attribution against adversarial post-processes.
|
true |
On the Transfer of Disentangled Representations in Realistic Settings
representation learning disentanglement real-world
Learning meaningful representations that disentangle the underlying structure of the data generating process is considered to be of key importance in machine learning. While disentangled representations were found to be useful for diverse tasks such as abstract reasoning and fair classification, their scalability and real-world impact remain questionable.
We introduce a new high-resolution dataset with 1M simulated images and over 1,800 annotated real-world images of the same setup. In contrast to previous work, this new dataset exhibits correlations, a complex underlying structure, and allows to evaluate transfer to unseen simulated and real-world settings where the encoder i) remains in distribution or ii) is out of distribution.
We propose new architectures in order to scale disentangled representation learning to realistic high-resolution settings and conduct a large-scale empirical study of disentangled representations on this dataset. We observe that disentanglement is a good predictor for out-of-distribution (OOD) task performance.
|
false |
Explainability for fair machine learning
explainability fairness Shapley
As the decisions made or influenced by machine learning models increasingly impact our lives, it is crucial to detect, understand, and mitigate unfairness. But even simply determining what ``unfairness'' should mean in a given context is non-trivial: there are many competing definitions, and choosing between them often requires a deep understanding of the underlying task. It is thus tempting to use model explainability to gain insights into model fairness, however existing explainability tools do not reliably indicate whether a model is indeed fair. In this work we present a new approach to explaining fairness in machine learning, based on the Shapley value paradigm. Our fairness explanations attribute a model's overall unfairness to individual input features, even in cases where the model does not operate on sensitive attributes directly. Moreover, motivated by the linearity of Shapley explainability, we propose a meta algorithm for applying existing training-time fairness interventions, wherein one trains a perturbation to the original model, rather than a new model entirely. By explaining the original model, the perturbation, and the fair-corrected model, we gain insight into the accuracy-fairness trade-off that is being made by the intervention. We further show that this meta algorithm enjoys both flexibility and stability benefits with no loss in performance.
|
false |
Signal Coding and Reconstruction using Spike Trains
spike trains signal encoding reconstruction kernel representer theorem compression convolutional matching pursuit COMP
In many animal sensory pathways, the transformation from external stimuli to spike trains is essentially deterministic. In this context, a new mathematical framework for coding and reconstruction, based on a biologically plausible model of the spiking neuron, is presented. The framework considers encoding of a signal through spike trains generated by an ensemble of neurons via a standard convolve-then-threshold mechanism, albeit with a wide variety of convolution kernels. Neurons are distinguished by their convolution kernels and threshold values. Reconstruction is posited as a convex optimization minimizing energy. Formal conditions under which perfect reconstruction of the signal from the spike trains is possible are then identified. Coding experiments on a large audio dataset are presented to demonstrate the strength of the framework.
|
true |
Efficient Generalized Spherical CNNs
efficient computer vision natural sciences analysis spherical data representations equivariance rotational symmetries
Many problems across computer vision and the natural sciences require the analysis of spherical data, for which representations may be learned efficiently by encoding equivariance to rotational symmetries. We present a generalized spherical CNN framework that encompasses various existing approaches and allows them to be leveraged alongside each other. The only existing non-linear spherical CNN layer that is strictly equivariant has complexity $\mathcal{O}(C^2L^5)$, where $C$ is a measure of representational capacity and $L$ the spherical harmonic bandlimit. Such a high computational cost often prohibits the use of strictly equivariant spherical CNNs. We develop two new strictly equivariant layers with reduced complexity $\mathcal{O}(CL^4)$ and $\mathcal{O}(CL^3 \log L)$, making larger, more expressive models computationally feasible. Moreover, we adopt efficient sampling theory to achieve further computational savings. We show that these developments allow the construction of more expressive hybrid models that achieve state-of-the-art accuracy and parameter efficiency on spherical benchmark problems.
|
false |
Deep dive into CoCon - A Self Supervised approach for Controlled Text Generation
Natural language processing Language modeling controlled text generation
Transformer-based language models ([1] Vaswani et.al, 2017) have stirred transfer-based learning in NLP and have improved the performance of several NLP tasks. The preliminary step involves pretraining a language model on a large amount of text on the web. Research on steering a pretrained language model to enable fine-grained control over the content and sentiment of output is still under active exploration and has great potential in various applications such as story generation, search engines, etc. This blog post discusses a paper which proposes a content conditioner when trained auto-regressively alongside a Large pretrained language model (like GPT-2) provides the capability to control text at a fine-grained level.
|
false |
Improving the accuracy of neural networks in analog computing-in-memory systems by a generalized quantization method
analog computing-in-memory quantization algorithm deep neural networks
Crossbar-enabled analog computing-in-memory (CACIM) systems can significantly improve the computation speed and energy efficiency of deep neural networks (DNNs). However, the transition of DNN from the digital systems to CACIM systems usually reduces its accuracy. The major issue is that the weights of DNN are stored and calculated directly on analog quantities in CACIM systems. The variation and programming overhead of the analog weight limit the precision.
Therefore, a suitable quantization algorithm is important when deploying a DNN into CACIM systems to obtain less accuracy loss. The analog weight has its unique advantages when doing quantization. Because there is no encoding and decoding process, the set of quanta will not affect the computing process. Therefore, a generalized quantization method that does not constrain the range of quanta and can obtain less quantization error will be effective in CACIM systems. For the first time, we introduced a generalized quantization method into CACIM systems and showed superior performance on a series of computer vision tasks, such as image classification, object detection, and semantic segmentation. Using the generalized quantization method, the DNN with 8-level analog weights can outperform the 32-bit networks. With fewer levels, the generalized quantization method can obtain less accuracy loss than other uniform quantization methods.
|
false |
Conditional Generative Modeling for De Novo Hierarchical Multi-Label Functional Protein Design
protein design conditional generative adversarial networks gene ontology hierarchical multi-label GO GAN
The availability of vast protein sequence information and rich functional annotations thereof has a large potential for protein design applications in biomedicine and synthetic biology. To this date, there exists no method for the general-purpose design of proteins without any prior knowledge about the protein of interest, such as costly and rare structure information or seed sequence fragments. However, the Gene Ontology (GO) database provides information about the hierarchical organisation of protein functions, and thus could inform generative models about the underlying complex sequence-function relationships, replacing the need for structural data. We therefore propose to use conditional generative adversarial networks (cGANs) on the task of fast de novo hierarchical multi-label protein design. We generate protein sequences exhibiting properties of a large set of molecular functions extracted from the GO database, using a single model and without any prior information. We shed light on efficient conditioning mechanisms and adapted network architectures thanks to a thorough hyperparameter selection process and analysis. We further provide statistically- and biologically-driven evaluation measures for generative models in the context of protein design to assess the quality of the generated sequences and facilitate progress in the field. We show that our proposed model, ProteoGAN, outperforms several baselines when designing proteins given a functional label and generates well-formed sequences.
|
true |
Learning iterative algorithms to solve PDEs.
PDE Solver Physics-Informed
In this work, we propose a new method to solve partial differential equations (PDEs). Taking inspiration from traditional numerical methods, we view approx- imating solutions to PDEs as an iterative algorithm, and propose to learn the it- erations from data. With respect to directly predicting the solution with a neural network, our approach has access to the PDE, having the potential to enhance the model’s ability to generalize across a variety of scenarios, such as differing PDE parameters, initial or boundary conditions. We instantiate this framework and empirically validate its effectiveness across several PDE-solving benchmarks, evaluating efficiency and generalization capabilities, and demonstrating its poten- tial for broader applicability.
|
true |
Prototypical Contrastive Learning of Unsupervised Representations
self-supervised learning unsupervised learning representation learning contrastive learning
This paper presents Prototypical Contrastive Learning (PCL), an unsupervised representation learning method that bridges contrastive learning with clustering. PCL not only learns low-level features for the task of instance discrimination, but more importantly, it implicitly encodes semantic structures of the data into the learned embedding space. Specifically, we introduce prototypes as latent variables to help find the maximum-likelihood estimation of the network parameters in an Expectation-Maximization framework. We iteratively perform E-step as finding the distribution of prototypes via clustering and M-step as optimizing the network via contrastive learning. We propose ProtoNCE loss, a generalized version of the InfoNCE loss for contrastive learning, which encourages representations to be closer to their assigned prototypes. PCL outperforms state-of-the-art instance-wise contrastive learning methods on multiple benchmarks with substantial improvement in low-resource transfer learning. Code and pretrained models are available at https://github.com/salesforce/PCL.
|
true |
TropEx: An Algorithm for Extracting Linear Terms in Deep Neural Networks
linear regions linear terms deep learning theory deep neural networks rectified linear unit relu network piecewise linear function tropical function
Deep neural networks with rectified linear (ReLU) activations are piecewise linear functions, where hyperplanes partition the input space into an astronomically high number of linear regions. Previous work focused on counting linear regions to measure the network's expressive power and on analyzing geometric properties of the hyperplane configurations. In contrast, we aim to understand the impact of the linear terms on network performance, by examining the information encoded in their coefficients. To this end, we derive TropEx, a nontrivial tropical algebra-inspired algorithm to systematically extract linear terms based on data. Applied to convolutional and fully-connected networks, our algorithm uncovers significant differences in how the different networks utilize linear regions for generalization. This underlines the importance of systematic linear term exploration, to better understand generalization in neural networks trained with complex data sets.
|
true |
WiSE-ALE: Wide Sample Estimator for Aggregate Latent Embedding
Generative models Latent variable models Variational inference Auto-encoder Representation learning
In this paper, we present a new generative model for learning latent embeddings. Compared to the classical generative process, where each observed data point is generated from an individual latent variable, our approach assumes a global latent variable to generate the whole set of observed data points. We then propose a learning objective that is derived as an approximation to a lower bound to the data log likelihood, leading to our algorithm, WiSE-ALE. Compared to the standard ELBO objective, where the variational posterior for each data point is encouraged to match the prior distribution, the WiSE-ALE objective matches the averaged posterior, over all samples, with the prior, allowing the sample-wise posterior distributions to have a wider range of acceptable embedding mean and variance and leading to better reconstruction quality in the auto-encoding process. Through various examples and comparison to other state-of-the-art VAE models, we demonstrate that WiSE-ALE has excellent information embedding properties, whilst still retaining the ability to learn a smooth, compact representation.
|
false |
Prediction of Enzyme Specificity using Protein Graph Convolutional Neural Networks
graph convolutional neural networks protease specificity proteins Rosetta energy function
Specific molecular recognition by proteins, for example, protease enzymes, is critical for maintaining the robustness of key life processes. The substrate specificity landscape of a protease enzyme comprises the set of all sequence motifs that are recognized/cut, or just as importantly, not recognized/cut by the enzyme. Current methods for predicting protease specificity landscapes rely on learning sequence patterns in experimentally derived data with a single enzyme, but are not robust to even small mutational changes. A comprehensive evaluation of specificity requires consideration of the three-dimensional structure and energetics of molecular interactions. In this work, we present a protein graph convolutional neural network (PGCN), which uses a physically intuitive, structure-based molecular interaction graph generated using the Rosetta energy function that describes the topology and energetic features, to determine substrate specificity. We use the PGCN to recapitulate and predict the specificity of the NS3/4 protease from the Hepatitic C virus. We compare our PGCN with previously used machine learning models and show that its performance in classification tasks is equivalent or better. Because PGCN is based on physical interactions, it is inherently more interpretable; determination of feature importance reveals key sub-graph patterns responsible for molecular recognition that are biochemically reasonable. The PGCN model also readily lends itself to the design of novel enzymes with tailored specificity against disease targets.
|
true |
FedMix: Approximation of Mixup under Mean Augmented Federated Learning
federated learning mixup
Federated learning (FL) allows edge devices to collectively learn a model without directly sharing data within each device, thus preserving privacy and eliminating the need to store data globally. While there are promising results under the assumption of independent and identically distributed (iid) local data, current state-of-the-art algorithms suffer a performance degradation as the heterogeneity of local data across clients increases. To resolve this issue, we propose a simple framework, \emph{Mean Augmented Federated Learning (MAFL)}, where clients send and receive \emph{averaged} local data, subject to the privacy requirements of target applications. Under our framework, we propose a new augmentation algorithm, named \emph{FedMix}, which is inspired by a phenomenal yet simple data augmentation method, Mixup, but does not require local raw data to be directly shared among devices. Our method shows greatly improved performance in the standard benchmark datasets of FL, under highly non-iid federated settings, compared to conventional algorithms.
|
true |
Split Batch Normalization: Improving Semi-Supervised Learning under Domain Shift
semi-supervised learning domain shift image classification deep neural networks
Recent work has shown that using unlabeled data in semi-supervised learning is not always beneficial and can even hurt generalization, especially when there is a class mismatch between the unlabeled and labeled examples. We investigate this phenomenon for image classification and many other forms of domain shifts (e.g. salt-and-pepper noise). Our main contribution is showing how to benefit from additional unlabeled data that comes from a shifted distribution in batch-normalized neural networks. We achieve it by simply using separate batch normalization statistics for unlabeled examples. Due to its simplicity, we recommend it as a standard practice.
|
true |
Learning to Sample with Local and Global Contexts in Experience Replay Buffer
reinforcement learning experience replay buffer off-policy RL
Experience replay, which enables the agents to remember and reuse experience from the past, has played a significant role in the success of off-policy reinforcement learning (RL). To utilize the experience replay efficiently, the existing sampling methods allow selecting out more meaningful experiences by imposing priorities on them based on certain metrics (e.g. TD-error). However, they may result in sampling highly biased, redundant transitions since they compute the sampling rate for each transition independently, without consideration of its importance in relation to other transitions. In this paper, we aim to address the issue by proposing a new learning-based sampling method that can compute the relative importance of transition. To this end, we design a novel permutation-equivariant neural architecture that takes contexts from not only features of each transition (local) but also those of others (global) as inputs. We validate our framework, which we refer to as Neural Experience Replay Sampler (NERS), on multiple benchmark tasks for both continuous and discrete control tasks and show that it can significantly improve the performance of various off-policy RL methods. Further analysis confirms that the improvements of the sample efficiency indeed are due to sampling diverse and meaningful transitions by NERS that considers both local and global contexts.
|
true |
Conversational grounding in emergent communication -- data and divergence
emergent communication conversational grounding data collection
We argue for a new research direction in emergent communication, combining work on `Conversational Grounding' with Symbol Grounding (SG+CG). We first present the fine-grained and targeted feedback signals provided by Conversational Grounding messages and discuss the potential advantages of such a combination. We argue that a key factor holding back research in this area is lack of appropriate data, where divergent agents can resolve disagreements and errors, and we propose requirements and methods for new data collections enabling such work.
|
true |
Learning to Coarsen Graphs with Graph Neural Networks
relational learning graph convolutions graph theory
With the rise of large-scale graphs for relational learning, graph coarsening emerges as a computationally viable alternative. We revisit the principles that aim to improve data-driven graph coarsening with adjustable coarsened structures.
|
false |
Benchmarking Multi-Agent Deep Reinforcement Learning Algorithms
Multi-agent Reinforcement Learning Benchmarking
We benchmark commonly used multi-agent deep reinforcement learning (MARL) algorithms on a variety of cooperative multi-agent games. While there has been significant innovation in MARL algorithms, algorithms tend to be tested and tuned on a single domain and their average performance across multiple domains is less characterized. Furthermore, since the hyperparameters of the algorithms are carefully tuned to the task of interest, it is unclear whether hyperparameters can easily be found that allow the algorithm to be repurposed for other cooperative tasks with different reward structure and environment dynamics. To investigate the consistency of the performance of MARL algorithms, we build an open-source library of multi-agent algorithms including DDPG/TD3/SAC with centralized Q functions, PPO with centralized value functions, and QMix and test them across a range of tasks that vary in coordination difficulty and agent number. The domains include the particle-world environments, starcraft micromanagement challenges, the Hanabi challenge, and the hide-and-seek environments. Finally, we investigate the ease of hyper-parameter tuning for each of the algorithms by tuning hyper-parameters in one environment per domain and re-using them in the other environments within the domain.
|
true |
Balancing Constraints and Rewards with Meta-Gradient D4PG
reinforcement learning meta-gradients constraints
Deploying Reinforcement Learning (RL) agents to solve real-world applications often requires satisfying complex system constraints. Often the constraint thresholds are incorrectly set due to the complex nature of a system or the inability to verify the thresholds offline (e.g, no simulator or reasonable offline evaluation procedure exists). This results in solutions where a task cannot be solved without violating the constraints. However, in many real-world cases, constraint violations are undesirable yet they are not catastrophic, motivating the need for soft-constrained RL approaches. We present two soft-constrained RL approaches that utilize meta-gradients to find a good trade-off between expected return and minimizing constraint violations. We demonstrate the effectiveness of these approaches by showing that they consistently outperform the baselines across four different Mujoco domains.
|
false |
Meta-Reinforcement Learning Robust to Distributional Shift via Model Identification and Experience Relabeling
Meta-Reinforcement Learning Meta Learning Reinforcement Learning
Reinforcement learning algorithms can acquire policies for complex tasks autonomously. However, the number of samples required to learn a diverse set of skills can be prohibitively large. While meta-reinforcement learning methods have enabled agents to leverage prior experience to adapt quickly to new tasks, their performance depends crucially on how close the new task is to the previously experienced tasks. Current approaches are either not able to extrapolate well, or can do so at the expense of requiring extremely large amounts of data for on-policy meta-training. In this work, we present model identification and experience relabeling (MIER), a meta-reinforcement learning algorithm that is both efficient and extrapolates well when faced with out-of-distribution tasks at test time. Our method is based on a simple insight: we recognize that dynamics models can be adapted efficiently and consistently with off-policy data, more easily than policies and value functions. These dynamics models can then be used to continue training policies and value functions for out-of-distribution tasks without using meta-reinforcement learning at all, by generating synthetic experience for the new task.
|
false |
Random Coordinate Langevin Monte Carlo
Random coordinate descent Langevin dynamics Monte Carlo sampling
Langevin Monte Carlo (LMC) is a popular Markov chain Monte Carlo sampling method. One drawback is that it requires the computation of the full gradient at each iteration, an expensive operation if the dimension of the problem is high. We propose a new sampling method: Random Coordinate LMC (RC-LMC). At each iteration, a single coordinate is randomly selected to be updated by a multiple of the partial derivative along this direction plus noise, and all other coordinates remain untouched. We investigate the total complexity of RC-LMC and compare it with the classical LMC for log-concave probability distributions. When the gradient of the log-density is Lipschitz, RC-LMC is less expensive than the classical LMC if the log-density is highly skewed for high dimensional problems, and when both the gradient and the Hessian of the log-density are Lipschitz, RC-LMC is always cheaper than the classical LMC, by a factor proportional to the square root of the problem dimension. In the latter case, our estimate of complexity is sharp with respect to the dimension.
|
true |
Early Stopping in Deep Networks: Double Descent and How to Eliminate it
early stopping double descent
Over-parameterized models, such as large deep networks, often exhibit a double descent phenomenon, whereas a function of model size, error first decreases, increases, and decreases at last. This intriguing double descent behavior also occurs as a function of training epochs and has been conjectured to arise because training epochs control the model complexity. In this paper, we show that such epoch-wise double descent occurs for a different reason: It is caused by a superposition of two or more bias-variance tradeoffs that arise because different parts of the network are learned at different epochs, and mitigating this by proper scaling of stepsizes can significantly improve the early stopping performance. We show this analytically for i) linear regression, where differently scaled features give rise to a superposition of bias-variance tradeoffs, and for ii) a wide two-layer neural network, where the first and second layers govern bias-variance tradeoffs. Inspired by this theory, we study two standard convolutional networks empirically and show that eliminating epoch-wise double descent through adjusting stepsizes of different layers improves the early stopping performance.
|
true |
Bayesian Few-Shot Classification with One-vs-Each Pólya-Gamma Augmented Gaussian Processes
few-shot learning gaussian processes bayesian deep learning uncertainty estimation
Few-shot classification (FSC), the task of adapting a classifier to unseen classes given a small labeled dataset, is an important step on the path toward human-like machine learning. Bayesian methods are well-suited to tackling the fundamental issue of overfitting in the few-shot scenario because they allow practitioners to specify prior beliefs and update those beliefs in light of observed data. Contemporary approaches to Bayesian few-shot classification maintain a posterior distribution over model parameters, which is slow and requires storage that scales with model size. Instead, we propose a Gaussian process classifier based on a novel combination of Pólya-Gamma augmentation and the one-vs-each softmax approximation that allows us to efficiently marginalize over functions rather than model parameters. We demonstrate improved accuracy and uncertainty quantification on both standard few-shot classification benchmarks and few-shot domain transfer tasks.
|
false |
Copilot Evaluation Harness: Building User Trust in LLMs and LM Agents for IDE Environments
Machine Learning LLMs for Code Generation Reliability Execution-Based Benchmark
The addition of Large Language Models (LLMs) into Integrated Code Development Environments (IDEs) has become a focal point in modern software development. LLMs offer the potential to significantly augment developer productivity by serving as intelligent, chat-driven programming assistants, especially with the increase in LLM-driven coding agents. With these tools comes the need for safeguards and metrics for quality assurance for consumers. In this paper, we introduce the Copilot Evaluation Harness: a set of data and tools for evaluating LLM-guided coding, covering various programming scenarios and languages. We propose a more robust system for measuring and understanding model behavior when leveraged as chat coding assistants or coding agents than previous state of the art evaluation metrics.
We design and compute both static and execution-based success metrics on a wide range of developer tasks, including documentation generation from code (doc), test case generation (test), and bug-fixing (fix). In the chat scenario, we see that GPT4o has much lower prompt sensitivity than the other models. In the agentic scenario, we find that reasoning models are more inclined to generate one-shot solutions, even when given multiple turns and access to tool calling. We show how results from our metrics can be used to increase the interpretability and explainability of LLMs in the real-world IDE-chat scenario.
|
true |
PRUNING AS A DEFENSE: REDUCING MEMORIZATION IN LARGE LANGUAGE MODELS
pruning memorization LLMs
Large language models have been shown to memorize significant portions of their training data, which they can reproduce when appropriately prompted. This work investigates the impact of simple pruning techniques on this behavior. Our findings reveal that pruning effectively reduces the extent of memorization in LLMs, demonstrating its potential as a foundational approach for mitigating membership inference attacks.
|
true |
Investigating the effects of plant diversity on soil thermal diffusivity using Physics- Informed Neural Networks
Soil Temperature Diffusity Plant Diversity PINNs
The influence of plant diversity on the stability of ecosystems is well-reported in the literature.
However, the exact mechanisms responsible for this effect are still a topic of debate.
Recently, soil temperature stability was proposed as one possible candidate for such a mechanism.
To further evaluate this hypothesis, we investigate the relationship between plant diversity and the thermal diffusivity of the soil during the very dry and hot summer of 2018 in Central Europe.
By leveraging Physics-Informed Neural Networks and a 30-minute resolution soil temperature dataset from the Jena Experiment, we find an inverse relationship between plant diversity and the thermal diffusivity of the associated soil.
With this, we provide support for the idea of plant diversity as a natural protection against climate-related ecosystem change.
|
true |
On the mapping between Hopfield networks and Restricted Boltzmann Machines
Hopfield Networks Restricted Boltzmann Machines Statistical Physics
Hopfield networks (HNs) and Restricted Boltzmann Machines (RBMs) are two important models at the interface of statistical physics, machine learning, and neuroscience. Recently, there has been interest in the relationship between HNs and RBMs, due to their similarity under the statistical mechanics formalism. An exact mapping between HNs and RBMs has been previously noted for the special case of orthogonal (“uncorrelated”) encoded patterns. We present here an exact mapping in the case of correlated pattern HNs, which are more broadly applicable to existing datasets. Specifically, we show that any HN with $N$ binary variables and $p<N$ potentially correlated binary patterns can be transformed into an RBM with $N$ binary visible variables and $p$ gaussian hidden variables. We outline the conditions under which the reverse mapping exists, and conduct experiments on the MNIST dataset which suggest the mapping provides a useful initialization to the RBM weights. We discuss extensions, the potential importance of this correspondence for the training of RBMs, and for understanding the performance of feature extraction methods which utilize RBMs.
|
true |
Extending Deep Learning Emulation Across Parameter Regimes to Assess Stochastically Driven Spontaneous Transition Events
Probabilistic Deep Learning Fine-tuning; Generalisation Stochastic Dynamics Fluid Dynamics
Given the computational expense associated with simultaneous multi-task learn- ing, we leverage fine-tuning to generalise a transformer-based neural network emulating a stochastic dynamical system across a range of parameters. Fine-tuning a neural network with a dataset containing a set of parameter values yields a 40-fold reduction in required training size compared to training ab initio training for each new parameter. This facilitates rapid adaptation of the deep learning model, which can be used subsequently across a large range of the parameter space or tailored to a specific regime of study. We demonstrate the model’s ability to capture the relevant behaviour even at interpolated parameter values not seen during training. Applied to a well-researched zonal jet system, the speed-up provided by the deep learning model over numerical integration and the ability to sample from the probabilistic model makes uncertainty quantification in the form of statistical study of rare events in the physical system computationally feasible. Our code is available at https://github.com/Ira-Shokar/Stochastic-Transformer.
|
false |
Accurate and fast detection of copy number variations from short-read whole-genome sequencing with deep convolutional neural network
copy number variation deep learning convolutional neural network computational biology DNA sequencing
A copy number variant (CNV) is a type of genetic mutation where a stretch of DNA is lost or duplicated once or multiple times. CNVs play important roles in the development of diseases and complex traits. CNV detection with short-read DNA sequencing technology is challenging because CNVs significantly vary in size and are similar to DNA sequencing artifacts. Many methods have been developed but still yield unsatisfactory results with high computational costs. Here, we propose CNV-Net, a novel approach for CNV detection using a six-layer convolutional neural network. We encode DNA sequencing information into RGB images and train the convolutional neural network with these images. The fitted convolutional neural network can then be used to predict CNVs from DNA sequencing data. We benchmark CNV-Net with two high-quality whole-genome sequencing datasets available from the Genome in a Bottle Consortium, considered as gold standard benchmarking datasets for CNV detection. We demonstrate that CNV-Net is more accurate and efficient in CNV detection than current tools.
|
true |
Automated Red Teaming with GOAT: the Generative Offensive Agent Tester
red teaming adversarial machine learning adversarial examples attacks on language models
Red teaming assess how large language models (LLMs) can produce content that violates norms, policies, and rules set forth during their safety training. However, most existing automated methods in literature are not representative of the way common users exploit the multi-turn conversational nature of AI models. While manual testing addresses this gap, it is an inefficient and often expensive process. To address these limitations, we introduce the Generative Offensive Agent Tester (GOAT), an automated agentic red teaming system that simulates plain language adversarial conversations while leveraging multiple adversarial prompting techniques to identify vuLnerabilities in LLMs. We instantiate GOAT with 7 red teaming attacks by prompting a general purpose model in a way that encourages reasoning through the choices of methods available, the current target model’s response, and the next steps. Our approach is designed to be extensible and efficient, allowing human testers to focus on exploring new areas of risk while automation covers the scaled adversarial stress-testing of known risk territory. We present the design and evaluation of GOAT, demonstrating its effectiveness in identifying vulnerabilities in state-of-the-art LLMs, with an ASR@10 of 96% against smaller models such as Llama 3.1 8B, and 91% against Llama 3.1 70B and 94% for GPT-4o when evaluated against larger models on the JailbreakBench dataset.
|
true |
ASIDE: Architectural Separation of Instructions and Data in Language Models
Instruction-data separation LLM Safety ML Safety Prompt Injections LLMs
Despite their remarkable performance, large language models lack elementary safety features, and this makes them susceptible to numerous malicious attacks. In particular, previous work has identified the absence of an intrinsic separation between instructions and data as a root cause for the success of prompt injection attacks. In this work, we propose an architectural change, ASIDE, that allows the model to clearly separate between instructions and data by using separate embeddings for them. Specifically, the data embedding is initialized with a rotation of the
pretrained model’s embedding, prompting the model to learn to treat instructions and data differently. We demonstrate the effectiveness of our method by showing (1) greatly increased instruction-data separation scores without a loss in model capabilities and (2) competitive results on prompt injection benchmarks, even without dedicated safety training. Additionally, we study the working mechanism behind our method through an analysis of model representations.
|
true |
Share or Not? Learning to Schedule Language-Specific Capacity for Multilingual Translation
language-specific modeling conditional computation multilingual translation multilingual transformer
Using a mix of shared and language-specific (LS) parameters has shown promise in multilingual neural machine translation (MNMT), but the question of when and where LS capacity matters most is still under-studied. We offer such a study by proposing conditional language-specific routing (CLSR). CLSR employs hard binary gates conditioned on token representations to dynamically select LS or shared paths. By manipulating these gates, it can schedule LS capacity across sub-layers in MNMT subject to the guidance of translation signals and budget constraints. Moreover, CLSR can easily scale up to massively multilingual settings. Experiments with Transformer on OPUS-100 and WMT datasets show that: 1) MNMT is sensitive to both the amount and the position of LS modeling: distributing 10%-30% LS computation to the top and/or bottom encoder/decoder layers delivers the best performance; and 2) one-to-many translation benefits more from CLSR compared to many-to-one translation, particularly with unbalanced training data. Our study further verifies the trade-off between the shared capacity and LS capacity for multilingual translation. We corroborate our analysis by confirming the soundness of our findings as foundation of our improved multilingual Transformers. Source code and models are available at https://github.com/bzhangGo/zero/tree/iclr2021_clsr.
|
false |
Universal Sentence Representations Learning with Conditional Masked Language Model
multilingual representations sentence embeddings
This paper presents a novel training method, Conditional Masked Language Modeling (CMLM), to effectively learn sentence representations on large scale unlabeled corpora. CMLM integrates sentence representation learning into MLM training by conditioning on the encoded vectors of adjacent sentences. Our English CMLM model achieves state-of-the-art performance on SentEval, even outperforming models learned using (semi-)supervised signals. As a fully unsupervised learning method, CMLM can be conveniently extended to a broad range of languages and domains. We find that a multilingual CMLM model co-trained with bitext retrieval~(BR) and natural language inference~(NLI) tasks outperforms the previous state-of-the-art multilingual models by a large margin. We explore the same language bias of the learned representations, and propose a principle component based approach to remove the language identifying information from the representation while still retaining sentence semantics.
|
false |
From Generalization to Inner Activations
generalization activations model side effects deep learning models large amount data researchers information
One of the side effects of deep learning models becoming increasingly large is the amount of data that they intermediately generate which is then often ignored. Researchers have used this information to create explainability methods which show importance relative to a specific input it relates to, but often the intermediate values are disregarded and mostly underutilized. Although this intermediate data can become increasingly unwieldily in size as the networks grow, the ability to monitor specific layers is a valuable tool that provides insight into how the model is learning as well as vague generalities about how the model performs overall (e.g. where are specific features begun to be extracted or where is noise filtered out).
|
false |
UNSUPERVISED ANOMALY DETECTION FROM SEMANTIC SIMILARITY SCORES
Anomaly Detection Out-of-Distribution Detection Novelty Detection
In this paper we present SemSAD, a simple and generic framework for detecting examples that lie out-of-distribution (OOD) for a given training set. The approach is based on learning a semantic similarity measure to find for a given test example the semantically closest example in the training set and then using a discriminator to classify whether the two examples show sufficient semantic dissimilarity such that the test example can be rejected as OOD. We are able to outperform previous approaches for anomaly, novelty, or out-of-distribution detection in the visual domain by a large margin. In particular we obtain AUROC values close to one for the challenging task of detecting examples from CIFAR-10 as out-of-distribution given CIFAR-100 as in-distribution, without making use of label information.
|
false |
Learning Representations with Seq2Seq Models for Damage Detection
Representation learning Seq2Seq model Damage detection
Natural hazards have incurred damages to buildings and economic losses worldwide. Post-hazard response requires an accurate and fast damage detection and assessment. The data-driven damage detection approach has been emerged as an alternative to the conventional human vision inspection. We use a Seq2Seq model to learn damage representations by training with only undamaged signals. We test the validity of our Seq2Seq model with a signal dataset that is collected from a 2-story timber building. Results show that our Seq2Seq model has a strong capability of distinguishing damage representations for different damage states. Our code is available at the repository: \url{https://github.com/qryang/Damage-representation}.
|
false |
Convergence Analysis of Homotopy-SGD for Non-Convex Optimization
deep learning numerical optimization transfer learning
First-order stochastic methods for solving large-scale non-convex optimization problems are widely used in many big-data applications, e.g. training deep neural networks as well as other complex and potentially non-convex machine learning
models. Their inexpensive iterations generally come together with slow global convergence rate (mostly sublinear), leading to the necessity of carrying out a very high number of iterations before the iterates reach a neighborhood of a minimizer. In this work, we present a first-order stochastic algorithm based on a combination of homotopy methods and SGD, called Homotopy-Stochastic Gradient Descent (H-SGD), which finds interesting connections with some proposed heuristics in the literature, e.g. optimization by Gaussian continuation, training by diffusion, mollifying networks. Under some mild and realistic assumptions on the problem structure, we conduct a theoretical analysis of the proposed algorithm. Our analysis shows that, with a specifically designed scheme for the homotopy parameter, H-SGD enjoys a global linear rate of convergence to a neighborhood of a minimizer while maintaining fast and inexpensive iterations. Experimental evaluations confirm the theoretical results and show that H-SGD can outperform standard SGD.
|
false |
Tight Second-Order Certificates for Randomized Smoothing
certificates adversarial robustness defenses smoothing curvature
Randomized smoothing is a popular way of providing robustness guarantees against adversarial attacks: randomly-smoothed functions have a universal Lipschitz-like bound, allowing for robustness certificates to be easily computed. In this work, we show that there also exists a universal curvature-like bound for Gaussian random smoothing: given the exact value and gradient of a smoothed function, we compute a lower bound on the distance of a point to its closest adversarial example, called the Second-order Smoothing (SoS) robustness certificate. In addition to proving the correctness of this novel certificate, we show that SoS certificates are realizable and therefore tight. Interestingly, we show that the maximum achievable benefits, in terms of certified robustness, from using the additional information of the gradient norm are relatively small: because our bounds are tight, this is a fundamental negative result. The gain of SoS certificates further diminishes if we consider the estimation error of the gradient norms, for which we have developed an estimator. We therefore additionally develop a variant of Gaussian smoothing, called Gaussian dipole smoothing, which provides similar bounds to randomized smoothing with gradient information, but with much-improved sample efficiency. This allows us to achieve (marginally) improved robustness certificates on high-dimensional datasets such as CIFAR-10 and ImageNet.
|
true |
Meta-Learning and Universality: Deep Representations and Gradient Descent can Approximate any Learning Algorithm
meta-learning learning to learn universal function approximation
Learning to learn is a powerful paradigm for enabling models to learn from data more effectively and efficiently. A popular approach to meta-learning is to train a recurrent model to read in a training dataset as input and output the parameters of a learned model, or output predictions for new test inputs. Alternatively, a more recent approach to meta-learning aims to acquire deep representations that can be effectively fine-tuned, via standard gradient descent, to new tasks. In this paper, we consider the meta-learning problem from the perspective of universality, formalizing the notion of learning algorithm approximation and comparing the expressive power of the aforementioned recurrent models to the more recent approaches that embed gradient descent into the meta-learner. In particular, we seek to answer the following question: does deep representation combined with standard gradient descent have sufficient capacity to approximate any learning algorithm? We find that this is indeed true, and further find, in our experiments, that gradient-based meta-learning consistently leads to learning strategies that generalize more widely compared to those represented by recurrent models.
|
true |
IDF++: Analyzing and Improving Integer Discrete Flows for Lossless Compression
normalizing flows lossless source compression generative modeling
In this paper we analyse and improve integer discrete flows for lossless compression. Integer discrete flows are a recently proposed class of models that learn invertible transformations for integer-valued random variables. Their discrete nature makes them particularly suitable for lossless compression with entropy coding schemes. We start by investigating a recent theoretical claim that states that invertible flows for discrete random variables are less flexible than their continuous counterparts. We demonstrate with a proof that this claim does not hold for integer discrete flows due to the embedding of data with finite support into the countably infinite integer lattice. Furthermore, we zoom in on the effect of gradient bias due to the straight-through estimator in integer discrete flows, and demonstrate that its influence is highly dependent on architecture choices and less prominent than previously thought. Finally, we show how different architecture modifications improve the performance of this model class for lossless compression, and that they also enable more efficient compression: a model with half the number of flow layers performs on par with or better than the original integer discrete flow model.
|
false |
Learning from multiscale wavelet superpixels using GNN with spatially heterogeneous pooling
Image classification GNN superpixel SLIC wavelet
Neural networks have become the standard for image classification tasks. On one hand, convolutional neural networks (CNNs) achieve state-of-the-art performance by learning from a regular grid representation of images. On the other hand, graph neural networks (GNNs) have shown promise in learning image classification from an embedded superpixel graph. However, in the latter, studies have been restricted to SLIC superpixels, where 1) a single target number of superpixels is arbitrarily defined for an entire dataset irrespective of differences across images and 2) the superpixels in a given image are of similar size despite intrinsic multiscale structure. In this study, we investigate learning from a new principled representation in which individual images are represented by an image-specific number of multiscale superpixels. We propose WaveMesh, a wavelet-based superpixeling algorithm, where the number and sizes of superpixels in an image are systematically computed based on the image content. We also present WavePool, a spatially heterogeneous pooling scheme tailored to WaveMesh superpixels. We study the feasibility of learning from the WaveMesh superpixel representation using SplineCNN, a state-of-the-art network for image graph classification. We show that under the same network architecture and training settings, SplineCNN with original Graclus-based pooling learns from WaveMesh superpixels on-par with SLIC superpixels. Additionally, we observe that the best performance is achieved when replacing Graclus-based pooling with WavePool while using WaveMesh superpixels.
|
false |
Variational saliency maps for explaining model's behavior
Interpretability XAI Variational Inference
Saliency maps have been widely used to explain the behavior of an image classifier. We introduce a new interpretability method which considers a saliency map as a random variable and aims to calculate the posterior distribution over the saliency map. The likelihood function is designed to measure the distance between the classifier's predictive probability of an image and that of locally perturbed image. For the prior distribution, we make attributions of adjacent pixels have a positive correlation. We use a variational approximation, and show that the approximate posterior is effective in explaining the classifier's behavior. It also has benefits of providing uncertainty over the explanation, giving auxiliary information to experts on how much the explanation is trustworthy.
|
false |
Calibrated Adversarial Refinement for Stochastic Semantic Segmentation
stochastic semantic segmentation conditional generative models adversarial training calibration uncertainty
Ambiguities in images or unsystematic annotation can lead to multiple valid solutions in semantic segmentation. To learn a distribution over predictions, recent work has explored the use of probabilistic networks. However, these do not necessarily capture the empirical distribution accurately. In this work, we aim to learn a calibrated multimodal predictive distribution, where the empirical frequency of the sampled predictions closely reflects that of the corresponding labels in the training set. To this end, we propose a novel two-stage, cascaded strategy for calibrated adversarial refinement. In the first stage, we explicitly model the data with a categorical likelihood. In the second, we train an adversarial network to sample from it an arbitrary number of coherent predictions. The model can be used independently or integrated into any black-box segmentation framework to facilitate learning of calibrated stochastic mappings. We demonstrate the utility and versatility of the approach by attaining state-of-the-art results on the multigrader LIDC dataset and a modified Cityscapes dataset. In addition, we use a toy regression dataset to show that our framework is not confined to semantic segmentation, and the core design can be adapted to other tasks requiring learning a calibrated predictive distribution.
|
false |
OpenCoS: Contrastive Semi-supervised Learning for Handling Open-set Unlabeled Data
semi-supervised learning realistic semi-supervised learning class-distribution mismatch unsupervised learning
Modern semi-supervised learning methods conventionally assume both labeled and unlabeled data have the same class distribution. However, unlabeled data may include out-of-class samples in practice; those that cannot have one-hot encoded labels from a closed-set of classes in label data, i.e., unlabeled data is an open-set. In this paper, we introduce OpenCoS, a method for handling this realistic semi-supervised learning scenario based on a recent framework of contrastive learning. One of our key findings is that out-of-class samples in the unlabeled dataset can be identified effectively via (unsupervised) contrastive learning. OpenCoS utilizes this information to overcome the failure modes in the existing state-of-the-art semi-supervised methods, e.g., ReMixMatch or FixMatch. In particular, we propose to assign soft-labels for out-of-class samples using the representation learned from contrastive learning. Our extensive experimental results show the effectiveness of OpenCoS, fixing the state-of-the-art semi-supervised methods to be suitable for diverse scenarios involving open-set unlabeled data.
|
true |
Causal-TGAN: Modeling Tabular Data Using Causally-Aware GAN
Tabular Data Generation Generative Adversarial Networks Causality
Generative adversarial net (GAN)-based tabular data generation has recently received significant attention for its power for data augmentation when available data is limited. Most prior works have applied generic GAN frameworks for tabular data generation without explicitly considering inter-variable relationships, which is important for modeling tabular data distribution. In this work, we design Causal-TGAN, a causally-aware generator architecture that can capture the relationships among variables (continuous-type, discrete-type, and mixed-type) by explicitly modeling the pre-defined inter-variable causal relationships. The flexibility of Causal-TGAN is its capability to support different degrees of subject matter expert domain knowledge (e.g., complete or partial) about the inter-variable causal relations. Extensive experimental results on both simulated and real-world datasets demonstrate that exploiting causal relations in deep generative models could improve the generated tabular data quality compared to the state-of-the-art. Code is available at \url{https://github.com/BiggyBing/Causal-TGAN-Public}.
|
false |
Understanding Mental Representations Of Objects Through Verbs Applied To Them
Affordance affordance embedding object representation
In order to interact with objects in our environment, we rely on an understanding of the actions that can be performed on them, and the extent to which they rely or have an effect on the properties of the object. This knowledge is called the object "affordance". We propose an approach for creating an embedding of objects in an affordance space, in which each dimension corresponds to an aspect of meaning shared by many actions, using text corpora. This embedding makes it possible to predict which verbs will be applicable to a given object, as captured in human judgments of affordance, better than a variety of alternative approaches. Furthermore, we show that the dimensions learned are interpretable, and that they correspond to typical patterns of interaction with objects. Finally, we show that the dimensions can be used to predict a state-of-the-art mental representation of objects, derived purely from human judgements of object similarity.
|
true |
Rethinking Hallucinations: Correctness, Consistency, and Prompt Multiplicity
hallucinations evaluation prompt sensitivity
Large language models (LLMs) are known to "hallucinate" by generating false or misleading outputs. Hallucinations pose various harms, from erosion of trust to widespread misinformation. Existing hallucination evaluation, however, focuses only on "correctness" and often overlooks "consistency", necessary to distinguish and address these harms. To bridge this gap, we introduce _prompt multiplicity_, a framework for quantifying consistency through prompt sensitivity. Our analysis reveals significant multiplicity (over 50% inconsistency in benchmarks like Med-HALT), suggesting that hallucination-related harms have been severely underestimated. Furthermore, we study the role of consistency in hallucination detection and mitigation. We find that: (a) detection techniques capture consistency, not correctness, and (b) mitigation techniques like RAG can introduce additional inconsistencies. By integrating prompt multiplicity into hallucination evaluation, we provide an improved framework of potential harms and uncover critical limitations in current detection and mitigation strategies.
|
true |
Bi-Directional Block Self-Attention for Fast and Memory-Efficient Sequence Modeling
deep learning attention mechanism sequence modeling natural language processing sentence embedding
Recurrent neural networks (RNN), convolutional neural networks (CNN) and self-attention networks (SAN) are commonly used to produce context-aware representations. RNN can capture long-range dependency but is hard to parallelize and not time-efficient. CNN focuses on local dependency but does not perform well on some tasks. SAN can model both such dependencies via highly parallelizable computation, but memory requirement grows rapidly in line with sequence length. In this paper, we propose a model, called "bi-directional block self-attention network (Bi-BloSAN)", for RNN/CNN-free sequence encoding. It requires as little memory as RNN but with all the merits of SAN. Bi-BloSAN splits the entire sequence into blocks, and applies an intra-block SAN to each block for modeling local context, then applies an inter-block SAN to the outputs for all blocks to capture long-range dependency. Thus, each SAN only needs to process a short sequence, and only a small amount of memory is required. Additionally, we use feature-level attention to handle the variation of contexts around the same word, and use forward/backward masks to encode temporal order information. On nine benchmark datasets for different NLP tasks, Bi-BloSAN achieves or improves upon state-of-the-art accuracy, and shows better efficiency-memory trade-off than existing RNN/CNN/SAN.
|
true |
A PHYSICS-INFORMED NEURAL NETWORK FOR COUPLED CALCIUM DYNAMICS IN A CABLE NEURON
physics-informed neural networks PINN transcranial magnetic stimulation TMS calcium dynamics neurphysiology
Transcranial magnetic stimulation (TMS) is a noninvasive treatment for a variety of neurological and neuropsychiatric disorders by triggering a calcium response through magnetic stimulation. To understand the full effects of this treatment, researchers will often use numerical simulations to model and study the calcium response. These simulations are limited to short-time simulations of single neurons due to computational complexity, restricting their use in clinical settings. In this paper, we explore an application of physics-informed neural networks (PINNs) to accurately produce long-time simulations of neuronal responses, opening the possibility of utilizing these methods in clinical applications to directly benefit patients.
|
false |
A Large-scale Study on Training Sample Memorization in Generative Modeling
GAN generative adversarial networks generative modeling memorization
Many recent developments on generative models for natural images have relied on heuristically-motivated metrics that can be easily gamed by memorizing a small sample from the true distribution or training a model directly to improve the metric.
In this work, we critically evaluate the gameability of the benchmarking procedure by running a competition which ultimately resulted in participants attempting to cheat. Our competition received over 11000 submitted models which allowed us to investigate memorization-aware metrics for measuring generative model performance. Specifically, we propose the Memorization-Informed Frechet Inception Distance (MiFID) and discuss ways to ensure that winning submissions were based on genuine improvements in perceptual quality. We evaluate the effectiveness of our benchmark by manually inspecting the code for the 1000 top-performing models and labeling different forms of memorization that were intentionally or unintentionally used. To facilitate future work on benchmarking generative models, we release generated images and our labels for these models as well as code to compute the MiFID metric.
|
true |
Improving Sample Complexity with Observational Supervision
observational supervision eye tracking gaze data limited labeled data
Supervised machine learning models for high-value computer vision applications such as medical image classification often require large datasets labeled by domain experts, which are slow to collect, expensive to maintain, and static with respect to changes in the data distribution. In this context, we assess the utility of observational supervision, where we take advantage of passively-collected signals such as eye tracking or “gaze” data, to reduce the amount of hand-labeled data needed for model training. Specifically, we leverage gaze information to directly supervise a visual attention layer by penalizing disagreement between the spatial regions the human labeler looked at the longest and those that most heavily influence model output. We present evidence that constraining the model in this way can reduce the number of labeled examples required to achieve a given performance level by as much as 50%, and that gaze information is most helpful on more difficult tasks.
|
true |
Gradient Vaccine: Investigating and Improving Multi-task Optimization in Massively Multilingual Models
Multi-task Learning Multilingual Modeling
Massively multilingual models subsuming tens or even hundreds of languages pose great challenges to multi-task optimization. While it is a common practice to apply a language-agnostic procedure optimizing a joint multilingual task objective, how to properly characterize and take advantage of its underlying problem structure for improving optimization efficiency remains under-explored. In this paper, we attempt to peek into the black-box of multilingual optimization through the lens of loss function geometry. We find that gradient similarity measured along the optimization trajectory is an important signal, which correlates well with not only language proximity but also the overall model performance. Such observation helps us to identify a critical limitation of existing gradient-based multi-task learning methods, and thus we derive a simple and scalable optimization procedure, named Gradient Vaccine, which encourages more geometrically aligned parameter updates for close tasks. Empirically, our method obtains significant model performance gains on multilingual machine translation and XTREME benchmark tasks for multilingual language models. Our work reveals the importance of properly measuring and utilizing language proximity in multilingual optimization, and has broader implications for multi-task learning beyond multilingual modeling.
|
true |
On Statistical Bias In Active Learning: How and When to Fix It
Active Learning Monte Carlo Risk Estimation
Active learning is a powerful tool when labelling data is expensive, but it introduces a bias because the training data no longer follows the population distribution. We formalize this bias and investigate the situations in which it can be harmful and sometimes even helpful. We further introduce novel corrective weights to remove bias when doing so is beneficial. Through this, our work not only provides a useful mechanism that can improve the active learning approach, but also an explanation for the empirical successes of various existing approaches which ignore this bias. In particular, we show that this bias can be actively helpful when training overparameterized models---like neural networks---with relatively modest dataset sizes.
|
true |
ToolScan: A Benchmark For Characterizing Errors In Tool-Use LLMs
LLM Function Calling Benchmark Error Detection Evaluation
Evaluating Large Language Models (LLMs) is one of the most critical aspects of building a performant compound AI system. Since the output from LLMs propagate to downstream steps, identifying LLM errors is crucial to system performance. A common task for LLMs in AI systems is tool use. While there are several benchmark environments for evaluating LLMs on this task, they typically only give a success rate without any explanation of the failure cases. To solve this problem, we introduce ToolScan, a new benchmark to identify error patterns in LLM output on tool-use tasks. Our benchmark data set comprises of queries from diverse environments that can be used to test for the presence of seven newly characterized error patterns. Using ToolScan, we show that even the most prominent LLMs exhibit these error patterns in their outputs. Researchers can use these insights from ToolScan to guide their error mitigation strategies. We open-source our evaluation framework at https://anonymous.4open.science/r/ToolScan-1474 .
|
false |
MC-LSTM: Mass-conserving LSTM
LSTM RNN mass-conservation neural arithmetic units inductive bias hydrology
The success of Convolutional Neural Networks (CNNs) in computer vision is mainly driven by their strong inductive bias, which is strong enough to allow CNNs to solve vision-related tasks with random weights, meaning without learning. Similarly, Long Short-Term Memory (LSTM) has a strong inductive bias towards storing information over time. However, many real-world systems are governed by conservation laws, which lead to the redistribution of particular quantities —e.g. in physical and economical systems. Our novel Mass-Conserving LSTM (MC-LSTM) adheres to these conservation laws by extending the inductive bias of LSTM to model the redistribution of those stored quantities. MC-LSTMs set a new state-of-the-art for neural arithmetic units at learning arithmetic operations, such as addition tasks, which have a strong conservation law, as the sum is constant overtime. Further, MC-LSTM is applied to traffic forecasting, modeling a pendulum, and a large benchmark dataset in hydrology, where it sets a new state-of-the-art for predicting peak flows. In the hydrology example, we show that MC-LSTM states correlate with real world processes and are therefore interpretable.
|
true |
Learning to Count Objects in Natural Images for Visual Question Answering
visual question answering vqa counting
Visual Question Answering (VQA) models have struggled with counting objects in natural images so far. We identify a fundamental problem due to soft attention in these models as a cause. To circumvent this problem, we propose a neural network component that allows robust counting from object proposals. Experiments on a toy task show the effectiveness of this component and we obtain state-of-the-art accuracy on the number category of the VQA v2 dataset without negatively affecting other categories, even outperforming ensemble models with our single model. On a difficult balanced pair metric, the component gives a substantial improvement in counting over a strong baseline by 6.6%.
|
true |
Multi-timescale Representation Learning in LSTM Language Models
Language Model LSTM timescales
Language models must capture statistical dependencies between words at timescales ranging from very short to very long. Earlier work has demonstrated that dependencies in natural language tend to decay with distance between words according to a power law. However, it is unclear how this knowledge can be used for analyzing or designing neural network language models. In this work, we derived a theory for how the memory gating mechanism in long short-term memory (LSTM) language models can capture power law decay. We found that unit timescales within an LSTM, which are determined by the forget gate bias, should follow an Inverse Gamma distribution. Experiments then showed that LSTM language models trained on natural English text learn to approximate this theoretical distribution. Further, we found that explicitly imposing the theoretical distribution upon the model during training yielded better language model perplexity overall, with particular improvements for predicting low-frequency (rare) words. Moreover, the explicit multi-timescale model selectively routes information about different types of words through units with different timescales, potentially improving model interpretability. These results demonstrate the importance of careful, theoretically-motivated analysis of memory and timescale in language models.
|
false |
Putting Theory to Work: From Learning Bounds to Meta-Learning Algorithms
meta-learning few-shot learning
Most of existing deep learning models rely on excessive amounts of labeled training data in order to achieve state-of-the-art results, even though these data can be hard or costly to get in practice. One attractive alternative is to learn with little supervision, commonly referred to as few-shot learning (FSL), and, in particular, meta-learning that learns to learn with few data from related tasks. Despite the practical success of meta-learning, many of its algorithmic solutions proposed in the literature are based on sound intuitions, but lack a solid theoretical analysis of the expected performance on the test task. In this paper, we review the recent advances in meta-learning theory and show how they can be used in practice both to better understand the behavior of popular meta-learning algorithms and to improve their generalization capacity. This latter is achieved by integrating the theoretical assumptions ensuring efficient meta-learning in the form of regularization terms into several popular meta-learning algorithms for which we provide a large study of their behavior on classic few-shot classification benchmarks. To the best of our knowledge, this is the first contribution that puts the most recent learning bounds of meta-learning theory into practice for the popular task of few-shot classification.
|
false |
XtraGPT: LLMs for Human-AI Collaboration on Controllable Scientific Paper Refinement
Paper Refinement Large Language Models
The increasing volume of scientific publications highlights the growing need for high-quality academic writing. However, while groundbreaking ideas are often present, many papers fail to meet academic writing standards. Unlike open-ended applications of large language models (LLMs) in research, which delegate creative tasks to AI, we emphasize a human-centered approach where researchers provide ideas and drafts while LLMs strictly follow user instructions for refinement. All XtraGPT data, training and evaluation processes, and models will be open-sourced.
We propose XtraGPT, LLMs designed to assist authors by delivering instruction-driven, context-aware revisions that (1) adhere to user instructions, (2) align with general academic writing standards, and (3) are consistent with the whole paper. Leveraging a dataset of 7,040 ICLR 24 papers and over 140,000 question-answer pairs, XtraGPT enhances specific sections without compromising the paper’s integrity. Experimental results show XtraGPT-7B surpass similar size models and is competitive with GPT-4o-mini in providing high-quality, context-aware refinements. We also found that scaling up model parameters provides limited improvement for the difficulty of paper scoring. Modifying six sections with XtraGPT can improve the paper’s rating according to the predictor.
By prioritizing controllability in the task of paper refinement, XtraGPT empowers researchers to focus on innovation while relying on the system to handle the demands of academic writing with context understanding and adherence to academic standards and user instructions.
|
false |
On Disentangled Representations Extracted from Pretrained GANs
Disentanglement representations GANs
Constructing disentangled representations is known to be a difficult task, especially in the unsupervised scenario. The dominating paradigm of unsupervised disentanglement is currently to train a generative model that separates different factors of variation in its latent space. This separation is typically enforced by training with specific regularization terms in the model's objective function. These terms, however, introduce additional hyperparameters responsible for the trade-off between disentanglement and generation quality. While tuning these hyperparameters is crucial for proper disentanglement, it is often unclear how to tune them without external supervision.
This paper investigates an alternative route to disentangled representations. Namely, we propose to extract such representations from the state-of-the-art GANs trained without disentangling terms in their objectives. This paradigm of post hoc disentanglement employs little or no hyperparameters when learning representations, while achieving results on par with existing state-of-the-art, as shown by comparison in terms of established disentanglement metrics, fairness, and the abstract reasoning task.
All our code and models are publicly available.
|
true |
Evaluation of a Robust Control System in Real-World Cable-Driven Parallel Robots
Reinforcement Learning DDPG PPO TRPO PID control cable-driven robot
This study evaluates the performance of classical and modern control methods for real-world Cable-Driven Parallel Robots (CDPRs), focusing on underconstrained systems with limited time discretization. A comparative analysis is conducted between classical PID controllers and modern reinforcement learning algorithms, including Deep Deterministic Policy Gradient (DDPG), Proximal Policy Optimization (PPO), Soft Actor-Critic (SAC), and Trust Region Policy Optimization (TRPO). The results demonstrate that TRPO outperforms other methods, achieving the lowest root mean square (RMS) errors across various trajectories and exhibiting robustness to larger time intervals between control updates. TRPO's ability to balance exploration and exploitation enables stable control in noisy, real-world environments, reducing reliance on high-frequency sensor feedback and computational demands. These findings highlight TRPO's potential as a robust solution for complex robotic control tasks, with implications for dynamic environments and future applications in sensor fusion or hybrid control strategies.
|
false |
A Simple Approach To Define Curricula For Training Neural Networks
Curriculum learning neural networks
In practice, sequence of mini-batches generated by uniform sampling of examples from the entire data is used for training neural networks. Curriculum learning is a training strategy that sorts the training examples by their difficulty and gradually exposes them to the learner. In this work, we propose two novel curriculum learning algorithms and empirically show their improvements in performance with convolutional and fully-connected neural networks on multiple real image datasets. Our dynamic curriculum learning algorithm tries to reduce the distance between the network weight and an optimal weight at any training step by greedily sampling examples with gradients that are directed towards the optimal weight. The curriculum ordering determined by our dynamic algorithm achieves a training speedup of $\sim 45\%$ in our experiments. We also introduce a new task-specific curriculum learning strategy that uses statistical measures such as standard deviation and entropy values to score the difficulty of data points in natural image datasets. We show that this new approach yields a mean training speedup of $\sim 43\%$ in the experiments we perform. Further, we also use our algorithms to learn why curriculum learning works. Based on our study, we argue that curriculum learning removes noisy examples from the initial phases of training, and gradually exposes them to the learner acting like a regularizer that helps in improving the generalization ability of the learner.
|
false |
Exploiting structured data for learning contagious diseases under incomplete testing
infectious diseases neural networks healthcare regularization structured data
One of the ways that machine learning algorithms can help control the spread of an infectious disease is by building models that predict who is likely to get infected whether or not they display any symptoms, making them good candidates for preemptive isolation. In this work we ask: can we build reliable infection prediction models when the observed data is collected under limited, and biased testing that prioritizes testing symptomatic individuals? Our analysis suggests that under favorable conditions, incomplete testing might be sufficient to achieve relatively good out-of-sample prediction error. Favorable conditions occur when untested-infected individuals have sufficiently different characteristics from untested-healthy, and when the infected individuals are "potent", meaning they infect a large majority of their neighbors. We develop an algorithm that predicts infections, and show that it outperforms benchmarks on simulated data. We apply our model to data from a large hospital to predict Clostridioides difficile infections; a communicable disease that is characterized by asymptomatic (i.e., untested) carriers. Using a proxy instead of the unobserved untested-infected state, we show that our model outperforms benchmarks in predicting infections.
|
true |
SALD: Sign Agnostic Learning with Derivatives
implicit neural representations 3D shapes learning sign agnostic learning
Learning 3D geometry directly from raw data, such as point clouds, triangle soups, or unoriented meshes is still a challenging task that feeds many downstream computer vision and graphics applications.
In this paper, we introduce SALD: a method for learning implicit neural representations of shapes directly from raw data. We generalize sign agnostic learning (SAL) to include derivatives: given an unsigned distance function to the input raw data, we advocate a novel sign agnostic regression loss, incorporating both pointwise values and gradients of the unsigned distance function. Optimizing this loss leads to a signed implicit function solution, the zero level set of which is a high quality and valid manifold approximation to the input 3D data. The motivation behind SALD is that incorporating derivatives in a regression loss leads to a lower sample complexity, and consequently better fitting. In addition, we provide empirical evidence, as well as theoretical motivation in 2D that SAL enjoys a minimal surface property, favoring minimal area solutions. More importantly, we are able to show that this property still holds for SALD, i.e., with derivatives included.
We demonstrate the efficacy of SALD for shape space learning on two challenging datasets: ShapeNet that contains inconsistent orientation and non-manifold meshes, and D-Faust that contains raw 3D scans (triangle soups). On both these datasets, we present state-of-the-art results.
|
false |
Learning disentangled representations with the Wasserstein Autoencoder
generative modeling disentangle learning wasserstein autoencoder
Disentangled representation learning has undoubtedly benefited from objective function surgery. However, a delicate balancing act of tuning is still required in order to trade off reconstruction fidelity versus disentanglement. Building on previous successes of penalizing the total correlation in the latent variables, we propose TCWAE (Total Correlation Wasserstein Autoencoder). Working in the WAE paradigm naturally enables the separation of the total-correlation term, thus providing disentanglement control over the learned representation, while offering more flexibility in the choice of reconstruction cost. We propose two variants using different KL estimators and perform extensive quantitative comparisons on data sets with known generative factors, showing competitive results relative to state-of-the-art techniques. We further study the trade off between disentanglement and reconstruction on more-difficult data sets with unknown generative factors, where we expect improved reconstructions due to the flexibility of the WAE paradigm.
|
true |
Learning continuous-time PDEs from sparse data with graph neural networks
dynamical systems partial differential equations PDEs graph neural networks continuous time
The behavior of many dynamical systems follow complex, yet still unknown partial differential equations (PDEs). While several machine learning methods have been proposed to learn PDEs directly from data, previous methods are limited to discrete-time approximations or make the limiting assumption of the observations arriving at regular grids. We propose a general continuous-time differential model for dynamical systems whose governing equations are parameterized by message passing graph neural networks. The model admits arbitrary space and time discretizations, which removes constraints on the locations of observation points and time intervals between the observations. The model is trained with continuous-time adjoint method enabling efficient neural PDE inference. We demonstrate the model's ability to work with unstructured grids, arbitrary time steps, and noisy observations. We compare our method with existing approaches on several well-known physical systems that involve first and higher-order PDEs with state-of-the-art predictive performance.
|
true |
CausalWorld: A Robotic Manipulation Benchmark for Causal Structure and Transfer Learning
reinforcement learning transfer learning sim2real transfer domain adaptation causality generalization robotics
Despite recent successes of reinforcement learning (RL), it remains a challenge for agents to transfer learned skills to related environments. To facilitate research addressing this problem, we proposeCausalWorld, a benchmark for causal structure and transfer learning in a robotic manipulation environment. The environment is a simulation of an open-source robotic platform, hence offering the possibility of sim-to-real transfer. Tasks consist of constructing 3D shapes from a set of blocks - inspired by how children learn to build complex structures. The key strength of CausalWorld is that it provides a combinatorial family of such tasks with common causal structure and underlying factors (including, e.g., robot and object masses, colors, sizes). The user (or the agent) may intervene on all causal variables, which allows for fine-grained control over how similar different tasks (or task distributions) are. One can thus easily define training and evaluation distributions of a desired difficulty level, targeting a specific form of generalization (e.g., only changes in appearance or object mass). Further, this common parametrization facilitates defining curricula by interpolating between an initial and a target task. While users may define their own task distributions, we present eight meaningful distributions as concrete benchmarks, ranging from simple to very challenging, all of which require long-horizon planning as well as precise low-level motor control. Finally, we provide baseline results for a subset of these tasks on distinct training curricula and corresponding evaluation protocols, verifying the feasibility of the tasks in this benchmark.
|
false |
Federated Learning With Quantized Global Model Updates
Federated learning lossy broadcasting
We study federated learning (FL), which enables mobile devices to utilize their local datasets to collaboratively train a global model with the help of a central server, while keeping data localized. At each iteration, the server broadcasts the current global model to the devices for local training, and aggregates the local model updates from the devices to update the global model. Previous work on the communication efficiency of FL has mainly focused on the aggregation of model updates from the devices, assuming perfect broadcasting of the global model. In this paper, we instead consider broadcasting a compressed version of the global model. This is to further reduce the communication cost of FL, which can be particularly limited when the global model is to be transmitted over a wireless medium. We introduce a lossy FL (LFL) algorithm, in which both the global model and the local model updates are quantized before being transmitted. We analyze the convergence behavior of the proposed LFL algorithm assuming the availability of accurate local model updates at the server. Numerical experiments show that the proposed LFL scheme, which quantizes the global model update (with respect to the global model estimate at the devices) rather than the global model itself, significantly outperforms other existing schemes studying quantization of the global model at the PS-to-device direction. Also, the performance loss of the proposed scheme is marginal compared to the fully lossless approach, where the PS and the devices transmit their messages entirely without any quantization.
|
false |
Hybrid and Non-Uniform DNN quantization methods using Retro Synthesis data for efficient inference
quantization dnn inference data free quantization synthetic data model compression
Existing post-training quantization methods attempt to compensate for the quantization loss by determining the quantized weights and activation ranges with the help of training data. Quantization aware training methods, on the other hand, achieve accuracy near to FP32 models by training the quantized model which consume more time. Both these methods are not effective for privacy constraint applications as they are tightly coupled with training data. In contrast, this paper proposes a data-independent post-training quantization scheme that eliminates the need for training data. This is achieved by generating a faux dataset hereafter called as $\textit{‘Retro-Synthesis Data’}$ from the FP32 model layer statistics and further using it for quantization. This approach outperformed state-of-the-art methods including, but not limited to, ZeroQ and DFQ on models with and without batch-normalization layers for 8, 6 and 4 bit precisions. We also introduced two futuristic variants of post-training quantization methods namely $\textit{‘Hybrid-Quantization’}$ and $\textit{‘Non-Uniform Quantization’}$. The Hybrid-Quantization scheme determines the sensitivity of each layer for per-tensor and per-channel quantization, and thereby generates hybrid quantized models that are $10 - 20\%$ efficient in inference time while achieving same or better accuracy as compared to per-channel quantization. Also this method outperformed FP32 accuracy when applied for models such as ResNet-18, and ResNet-50 onImageNet dataset. In the proposed Non-Uniform quantization scheme, the weights are grouped into different clusters and these clusters are assigned with a varied number of quantization steps depending on the number of weights and their ranges in respective cluster. This method resulted in an accuracy improvement of $1\%$ against state-of-the-art quantization methods on ImageNet dataset.
|
false |
Network-Agnostic Knowledge Transfer for Medical Image Segmentation
Knowledge Transfer Deep Learning Medical Image Segmentation Pseudo Annotation
Conventional transfer learning leverages weights of pre-trained networks, but mandates the need for similar neural architectures. Alternatively, knowledge distillation can transfer knowledge between heterogeneous networks but often requires access to the original training data or additional generative networks. Knowledge transfer between networks can be improved by being agnostic to the choice of network architecture and reducing the dependence on original training data. We propose a knowledge transfer approach from a teacher to a student network wherein we train the student on an independent transferal dataset, whose annotations are generated by the teacher. Experiments were conducted on five state-of-the-art networks for semantic segmentation and seven datasets across three imaging modalities. We studied knowledge transfer from a single teacher, combination of knowledge transfer and fine-tuning, and knowledge transfer from multiple teachers. The student model with a single teacher achieved similar performance as the teacher; and the student model with multiple teachers achieved better performance than the teachers. The salient features of our algorithm include: 1) no need for original training data or generative networks, 2) knowledge transfer between different architectures, 3) ease of implementation for downstream tasks by using the downstream task dataset as the transferal dataset, 4) knowledge transfer of an ensemble of models, trained independently, into one student model. Extensive experiments demonstrate that the proposed algorithm is effective for knowledge transfer and easily tunable.
|
false |
Truly Deterministic Policy Optimization
Deterministic Policy Gradient Deterministic Exploration Reinforcement Learning
In this paper, we present a policy gradient method that avoids exploratory noise injection and performs policy search over the deterministic landscape. By avoiding noise injection all sources of estimation variance can be eliminated in systems with deterministic dynamics (up to the initial state distribution). Since deterministic policy regularization is impossible using traditional non-metric measures such as the KL divergence, we derive a Wasserstein-based quadratic model for our purposes. We state conditions on the system model under which it is possible to establish a monotonic policy improvement guarantee, propose a surrogate function for policy gradient estimation, and show that it is possible to compute exact advantage estimates if both the state transition model and the policy are deterministic. Finally, we describe two novel robotic control environments---one with non-local rewards in the frequency domain and the other with a long horizon (8000 time-steps)---for which our policy gradient method (TDPO) significantly outperforms existing methods (PPO, TRPO, DDPG, and TD3).
|
true |
Learning a vector field from snapshots of unidentified particles rather than particle trajectories
Stochastic differential equations Schrödinger bridges Dynamical systems
Practitioners frequently aim to infer dynamical system behaviors using snapshots at certain time points. For instance, in single-cell sequencing, to sequence a cell we must destroy it. So we cannot access a full trajectory of the behaviors of a cell, but we can access a snapshot sample. While stochastic differential equations (SDEs) are commonly used to analyze systems with full trajectory access, the availability of only sparse time samples without individual trajectory data makes traditional SDE learning methods inapplicable. Recent works in the deep learning community have explored using Schrödinger bridges for dynamics estimation from such data. However, these methods are primarily tailored for interpolating between two time points and struggle when asked to infer the underlying dynamics that generate all observed data from multiple snapshots. In particular, a naive extension to multiple points performs piecewise perfect interpolation with- out considering the collective information from all snapshots. In contrast, we propose a new method that leverages an iterative projection mechanism inspired by Schrödinger bridges. Our method does not require that the inferred dynamics precisely match every snapshot, offering a substantial advantage in practical applications where perfect data alignment is rare. By incorporating information from the entirety of the dataset, our model provides a more robust and flexible frame- work for dynamics inference. We test our method using well-known simulated parametric models from systems biology and ecology.
|
true |
Bidirectional Variational Inference for Non-Autoregressive Text-to-Speech
text-to-speech speech synthesis non-autoregressive VAE
Although early text-to-speech (TTS) models such as Tacotron 2 have succeeded in generating human-like speech, their autoregressive architectures have several limitations: (1) They require a lot of time to generate a mel-spectrogram consisting of hundreds of steps. (2) The autoregressive speech generation shows a lack of robustness due to its error propagation property. In this paper, we propose a novel non-autoregressive TTS model called BVAE-TTS, which eliminates the architectural limitations and generates a mel-spectrogram in parallel. BVAE-TTS adopts a bidirectional-inference variational autoencoder (BVAE) that learns hierarchical latent representations using both bottom-up and top-down paths to increase its expressiveness. To apply BVAE to TTS, we design our model to utilize text information via an attention mechanism. By using attention maps that BVAE-TTS generates, we train a duration predictor so that the model uses the predicted duration of each phoneme at inference. In experiments conducted on LJSpeech dataset, we show that our model generates a mel-spectrogram 27 times faster than Tacotron 2 with similar speech quality. Furthermore, our BVAE-TTS outperforms Glow-TTS, which is one of the state-of-the-art non-autoregressive TTS models, in terms of both speech quality and inference speed while having 58% fewer parameters.
|
true |
Online Adversarial Purification based on Self-supervised Learning
Adversarial Robustness Self-Supervised Learning
Deep neural networks are known to be vulnerable to adversarial examples, where a perturbation in the input space leads to an amplified shift in the latent network representation. In this paper, we combine canonical supervised learning with self-supervised representation learning, and present Self-supervised Online Adversarial Purification (SOAP), a novel defense strategy that uses a self-supervised loss to purify adversarial examples at test-time. Our approach leverages the label-independent nature of self-supervised signals and counters the adversarial perturbation with respect to the self-supervised tasks. SOAP yields competitive robust accuracy against state-of-the-art adversarial training and purification methods, with considerably less training complexity. In addition, our approach is robust even when adversaries are given the knowledge of the purification defense strategy. To the best of our knowledge, our paper is the first that generalizes the idea of using self-supervised signals to perform online test-time purification.
|
false |
A Benchmark for Scalable Oversight Mechanisms
scalable oversight debate. alignment
As AI agents surpass human capabilities, scalable oversight -- the problem of effectively supplying human feedback to potentially superhuman AI models -- becomes increasingly critical to ensure alignment. While numerous scalable oversight mechanisms have been proposed, they lack a systematic empirical framework to evaluate and compare them. While recent works have tried to empirically study scalable oversight mechanisms -- particularly Debate -- we argue that they contain methodological flaws that limit their usefulness to AI alignment. We introduce the scalable oversight benchmark, a principled framework for evaluating human feedback mechanisms based on our agent score difference (ASD) metric, a measure of how effectively a mechanism advantages truth-telling over deception. We supply a Python package to facilitate rapid and competitive evaluation of scalable oversight protocols on our benchmark, and conduct some a demonstrative experiment benchmarking Debate.
|
false |
Transformers are Deep Infinite-Dimensional Non-Mercer Binary Kernel Machines
Transformer models attention models kernel methods reproducing kernel Banach spaces
Despite their ubiquity in core AI fields like natural language processing, the mechanics of deep attention-based neural networks like the ``Transformer'' model are not fully understood. In this article, we present a new perspective towards understanding how Transformers work. In particular, we show that the ``"dot-product attention" that is the core of the Transformer's operation can be characterized as a kernel learning method on a pair of Banach spaces. In particular, the Transformer's kernel is characterized as having an infinite feature dimension. Along the way we generalize the standard kernel learning problem to what we term a "binary" kernel learning problem, where data come from two input domains and a response is defined for every cross-domain pair. We prove a new representer theorem for these binary kernel machines with non-Mercer (indefinite, asymmetric) kernels (implying that the functions learned are elements of reproducing kernel Banach spaces rather than Hilbert spaces), and also prove a new universal approximation theorem showing that the Transformer calculation can learn any binary non-Mercer reproducing kernel Banach space pair. We experiment with new kernels in Transformers, and obtain results that suggest the infinite dimensionality of the standard Transformer kernel is partially responsible for its performance. This paper's results provide a new theoretical understanding of a very important but poorly understood model in modern machine learning.
|
false |
Hidden Markov models are recurrent neural networks: A disease progression modeling application
hidden markov models recurrent neural networks disease progression
Hidden Markov models (HMMs) are commonly used for disease progression modeling when the true state of a patient is not fully known. Since HMMs may have multiple local optima, performance can be improved by incorporating additional patient covariates to inform parameter estimation. To allow for this, we formulate a special case of recurrent neural networks (RNNs), which we name hidden Markov recurrent neural networks (HMRNNs), and prove that each HMRNN has the same likelihood function as a corresponding discrete-observation HMM. As a neural network, the HMRNN can also be combined with any other predictive neural networks that take patient covariate information as input. We first show that parameter estimates from HMRNNs are numerically close to those obtained from HMMs via the Baum-Welch algorithm, thus empirically validating their theoretical equivalence. We then demonstrate how the HMRNN can be combined with other neural networks to improve parameter estimation and prediction, using an Alzheimer's disease dataset. The HMRNN yields parameter estimates that improve disease forecasting performance and offer a novel clinical interpretation compared with a standard HMM.
|
false |
A Provably Convergent and Practical Algorithm for Min-Max Optimization with Applications to GANs
min-max optimization GANs
We present a first-order algorithm for nonconvex-nonconcave min-max optimization problems such as those that arise in training GANs. Our algorithm provably converges in $\mathrm{poly}(d,L, b)$ steps for any loss function $f:\mathbb{R}^d \times \mathbb{R}^d \rightarrow \mathbb{R}$ which is $b$-bounded with ${L}$-Lipschitz gradient. To achieve convergence, we 1) give a novel approximation to the global strategy of the max-player based on first-order algorithms such as gradient ascent, and 2) empower the min-player to look ahead and simulate the max-player’s response for arbitrarily many steps, but restrict the min-player to move according to updates sampled from a stochastic gradient oracle. Our algorithm, when used to train GANs on synthetic and real-world datasets, does not cycle, results in GANs that seem to avoid mode collapse, and achieves a training time per iteration and memory requirement similar to gradient descent-ascent.
|
false |
A priori guarantees of finite-time convergence for Deep Neural Networks
priori guarantees convergence deep neural networks analysis loss function lyapunov priori upper bound settling time previous studies deep
In this paper, we perform Lyapunov based analysis of the loss function to derive an a priori upper bound on the settling time of deep neural networks. While previous studies have attempted to understand deep learning using control theory framework, there is limited work on a priori finite time convergence analysis. Drawing from the advances in analysis of finite-time control of non-linear systems, we provide a priori guarantees of finite-time convergence in a deterministic control theoretic setting. We formulate the supervised learning framework as a control problem where weights of the network are control inputs and learning translates into a tracking problem. An analytical formula for finite-time upper bound on settling time is provided a priori under the assumptions of boundedness of input. Finally, we prove that our loss function is robust against input perturbations.
|
true |
ML-Bench: Evaluating Large Language Models and Agents for Machine Learning Tasks on Repository-Level Code
LLMs Code Generation Agents
Despite Large Language Models (LLMs) achieving impressive results in code generation, significant challenges remain in automated ML development, particularly in utilizing existing ML repositories effectively. Also, recently, people have developed LLM agents that attempt to interact with repository code (e.g., resolving issues), prompting the need for end-to-end evaluations starting from environment setup to deploying the repository rather than merely generating code in already-configured environments. These two gaps have motivated our development of ML-Bench, a benchmark rooted in real-world ML applications that leverage existing code repositories. ML-Bench encompasses annotated 9,641 examples across 18 GitHub repositories, challenging LLMs to accommodate user-specified arguments and documentation intricacies effectively. To evaluate both LLMs and agents, two setups are employed: ML-Bench-L for assessing LLMs' text-to-code conversion within a predefined deployment environment, and ML-Bench-A for testing autonomous agents in an end-to-end task execution within a Linux sandbox environment. Our findings indicate that while GPT-4o leads with a Pass@5 rate surpassing 50%, there remains significant scope for improvement, highlighted by issues such as hallucinated outputs and difficulties with bash script generation. Notably, in the more demanding ML-Agent-Bench, GPT-4o achieves a 76.47% success rate, reflecting the efficacy of iterative action and feedback in complex task resolution. Our resources, including code, data, and models, are available at \url{https://anonymous.4open.science/r/ML-Bench}.
|
false |
Emergent Properties of Foveated Perceptual Systems
Hybrid Perceptual Systems Foveation Visual Crowding Texture Two-stage models
We introduce foveated perceptual systems -- a hybrid architecture inspired by human vision, to explore the role of a \textit{texture-based} foveation stage on the nature and robustness of subsequently learned visual representation in machines. Specifically, these two-stage perceptual systems first foveate an image, inducing a texture-like encoding of peripheral information -- mimicking the effects of \textit{visual crowding} -- which is then relayed through a convolutional neural network (CNN) trained to perform scene categorization. We find that these foveated perceptual systems learn a visual representation that is \textit{distinct} from their non-foveated counterpart through experiments that probe: 1) i.i.d and o.o.d generalization; 2) robustness to occlusion; 3) a center image bias; and 4) high spatial frequency sensitivity. In addition, we examined the impact of this foveation transform with respect to two additional models derived with a rate-distortion optimization procedure to compute matched-resource systems: a lower resolution non-foveated system, and a foveated system with adaptive Gaussian blurring. The properties of greater i.i.d generalization, high spatial frequency sensitivity, and robustness to occlusion emerged exclusively in our foveated texture-based models, independent of network architecture and learning dynamics. Altogether, these results demonstrate that foveation -- via peripheral texture-based computations -- yields a distinct and robust representational format of scene information relative to standard machine vision approaches, and also provides symbiotic computational support that texture-based peripheral encoding has important representational consequences for processing in the human visual system.
|
false |
A Near-Optimal Recipe for Debiasing Trained Machine Learning Models
Fairness Classification Statistical Parity Deep Learning
We present an efficient and scalable algorithm for debiasing trained models, including deep neural networks (DNNs), which we prove to be near-optimal by bounding its excess Bayes risk. Unlike previous black-box reduction methods to cost-sensitive classification rules, the proposed algorithm operates on models that have been trained without having to retrain the model. Furthermore, as the algorithm is based on projected stochastic gradient descent (SGD), it is particularly attractive for deep learning applications. We empirically validate the proposed algorithm on standard benchmark datasets across both classical algorithms and modern DNN architectures and demonstrate that it outperforms previous post-processing approaches for unbiased classification.
|
true |
Generating Molecules via Chemical Reactions
Molecules Deep generative model of molecules molecule optimization deep generative model of graphs
Over the last few years exciting work in deep generative models has produced models able to suggest new organic molecules by generating strings, trees, and graphs representing their structure. While such models are able to generate molecules with desirable properties, their utility in practice is limited due to the difficulty in knowing how to synthesize these molecules. We therefore propose a new molecule generation model, mirroring a more realistic real-world process, where reactants are selected and combined to form more complex molecules. More specifically, our generative model proposes a bag of initial reactants (selected from a pool of commercially-available molecules) and uses a reaction model to predict how they react together to generate new molecules. Modeling the entire process of constructing a molecule during generation offers a number of advantages. First, we show that such a model has the ability to generate a wide, diverse set of valid and unique molecules due to the useful inductive biases of modeling reactions. Second, modeling synthesis routes rather than final molecules offers practical advantages to chemists who are not only interested in new molecules but also suggestions on stable and safe synthetic routes. Third, we demonstrate the capabilities of our model to also solve one-step retrosynthesis problems, predicting a set of reactants that can produce a target product.
|
false |
Deep Learning Optimization Theory - Trajectory Analysis of Gradient Descent
deep learning theory trajectory analysis neural tangent kernel
In recent years an obvious yet mysterious fact that stood across various experiments is the ability of gradient descent, a relatively simple first-order optimization method, to optimize an enormous number of parameters on highly non-convex loss functions. In some sense, this practical observation stands in contrast to classical statistical learning theory. This post will discuss the significant progress researchers are making in bridging this theory gap and demystifying gradient descent.
|
false |
Enhanced Semantic Alignment in Transformer Tracking via Position Learning and Force-Directed Attention
Transformer Tracking; Single Object Tracking; Semantic Alignment; Self-supervised Position Loss; Force-Directed Attention
In the field of visual object tracking, one-stream pipelines have become the mainstream framework due to its efficient integration of feature extraction and relationship modeling.
However, existing methods still face the issue of semantic misalignment:
firstly, the interaction of positional encoding between the two branches leads to a misalignment between feature semantics and position encoding;
secondly, traditional attention mechanisms fail to distinguish between semantic attraction and repulsion among features, resulting in semantic misalignment when the model processes these features.
To address these issues, we propose an Enhanced Semantic Alignment Transformer Tracker (ESAT) with position encode learning and force-directed attention mechanism.
By leveraging positional encoding loss, ESAT separately learns the absolute positional encodings of the target and search branches, distinguishing the locations of various tokens and their positive or negative relationships, thereby enhancing the semantic consistency between position and features.
Additionally, it incorporates a repulsion-attraction mechanism applied to the self-attention module, simulating dynamic interactions between nodes to improve feature discrimination.
Extensive experiments on multiple public tracking datasets show that our method outperforms many pipelines and achieves superior performance on five challenging benchmarks.
|
false |
Task-similarity Aware Meta-learning through Nonparametric Kernel Regression
Task-similarity Meta-learning Kernel regression Nonparametric regression Task-descriptors
This paper investigates the use of nonparametric kernel-regression to obtain a task- similarity aware meta-learning algorithm. Our hypothesis is that the use of task- similarity helps meta-learning when the available tasks are limited and may contain outlier/ dissimilar tasks. While existing meta-learning approaches implicitly assume the tasks as being similar, it is generally unclear how this task-similarity could be quantified and used in the learning. As a result, most popular meta- learning approaches do not actively use the similarity/dissimilarity between the tasks, but rely on availability of huge number of tasks for their working. Our contribution is a novel framework for meta-learning that explicitly uses task-similarity in the form of kernels and an associated meta-learning algorithm. We model the task-specific parameters to belong to a reproducing kernel Hilbert space where the kernel function captures the similarity across tasks. The proposed algorithm iteratively learns a meta-parameter which is used to assign a task-specific descriptor for every task. The task descriptors are then used to quantify the task-similarity through the kernel function. We show how our approach conceptually generalizes the popular meta-learning approaches of model-agnostic meta-learning (MAML) and Meta-stochastic gradient descent (Meta-SGD) approaches. Numerical experiments with regression and classification tasks show that our algorithm outperforms these approaches when the number of tasks is limited, even in the presence of out- lier or dissimilar tasks. This supports our hypothesis that task-similarity helps improve the meta-learning performance in task-limited and adverse settings.
|
false |
Emotional Manipulation is All You Need: A Framework for Evaluating Healthcare Misinformation in LLMs
Medical AI safety Prompt Injection Attacks Emotional manipulation in LLMs LLM jailbreak vulnerabilities Healthcare-specific adversarial attacks Trust and safety in medical LLMs LLM jailbreak vulnerabilities
Warning: This paper discusses potentially harmful healthcare misinformation patterns and LLM vulnerabilities
The integration of Large Language Models (LLMs) into healthcare applications has raised critical concerns about their susceptibility to generating harmful medical misinformation, particularly when faced with emotionally manipulated prompt injection attacks. Through systematic evaluation of 112 attack scenarios in eight state-of-the-art LLMs, we reveal that emotional manipulation coupled with prompt injection can increase the generation of dangerous medical misinformation without warning from a baseline of 6.2% to 37.5%. We also found that emotional content not only amplifies attack success rates, but also leads to more severe forms of misinformation. Notably, models vary widely in their susceptibility - while some models like Claude 3.5 Sonnet demonstrate strong resistance across all attack types, others show high vulnerability to emotional manipulation. These findings not only underscore critical vulnerabilities in LLM safety filters, but also emphasize the urgent need for enhanced protection against emotionally manipulated prompt injection attacks. Given that 39\% of the US population already believe in alternative cancer treatments, our research highlights the life-threatening implications of AI-generated health misinformation and provides crucial insights for developing more robust safety mechanisms before deployment in clinical settings.
|
true |
Disentangling Sequence Memorization and General Capability in Large Language Models
Memorization Unlearning Localization
Verbatim memorization in large language models remains a persistent and unsolved challenge, raising critical concerns for privacy, copyright, and responsible deployment. Existing research suggests that effective unlearning requires targeting the specific neurons responsible for memorization, as broad model updates fail to erase content reliably. However, we show that even these approaches rest on a flawed premise. Through controlled experiments, we demonstrate that memorized sequences are not naturally isolated to specific neurons during training, except in cases where the sequences are highly atypical. In this work, we put forward a new training paradigm that attempts to \textbf{isolate memorization to specific neurons by design}. The core challenge is that gradients from the repeated sequences entangle both ``generalizing'' features that improve general capability, in addition to sequence-specific memorization. We show that a simple change to standard training can implicitly disentangle these by leveraging metadata that identifies repeated sequences. We verify the efficacy of our method (\seqtd) in a proof-of-concept natural language setting and unveil the mechanism by which this disentanglement is possible through the training dynamics of memorization. We conclude by discussing the practical considerations of the deployment of \seqtd and highlight potential avenues for incorporating it into large-scale settings.
|
true |
Lossless Compression of Structured Convolutional Models via Lifting
weight sharing graph neural networks lifted inference relational learning dynamic computation graphs convolutional models
Lifting is an efficient technique to scale up graphical models generalized to relational domains by exploiting the underlying symmetries. Concurrently, neural models are continuously expanding from grid-like tensor data into structured representations, such as various attributed graphs and relational databases. To address the irregular structure of the data, the models typically extrapolate on the idea of convolution, effectively introducing parameter sharing in their, dynamically unfolded, computation graphs. The computation graphs themselves then reflect the symmetries of the underlying data, similarly to the lifted graphical models. Inspired by lifting, we introduce a simple and efficient technique to detect the symmetries and compress the neural models without loss of any information. We demonstrate through experiments that such compression can lead to significant speedups of structured convolutional models, such as various Graph Neural Networks, across various tasks, such as molecule classification and knowledge-base completion.
|
false |
Adversarial Masking: Towards Understanding Robustness Trade-off for Generalization
Adversarial Machine Learning Adversarial Robustness Adversarial Training Generalization
Adversarial training is a commonly used technique to improve model robustness against adversarial examples. Despite its success as a defense mechanism, adversarial training often fails to generalize well to unperturbed test data. While previous work assumes it is caused by the discrepancy between robust and non-robust features, in this paper, we introduce \emph{Adversarial Masking}, a new hypothesis that this trade-off is caused by different feature maskings applied. Specifically, the rescaling operation in the batch normalization layer, when combined together with ReLU activation, serves as a feature masking layer to select different features for model training. By carefully manipulating different maskings, a well-balanced trade-off can be achieved between model performance on unperturbed and perturbed data. Built upon this hypothesis, we further propose Robust Masking (RobMask), which constructs unique masking for every specific attack perturbation by learning a set of primary adversarial feature maskings. By incorporating different feature maps after the masking, we can distill better features to help model generalization. Sufficiently, adversarial training can be treated as an effective regularizer to achieve better generalization. Experiments on multiple benchmarks demonstrate that RobMask achieves significant improvement on clean test accuracy compared to strong state-of-the-art baselines.
|
false |
Evaluating representations by the complexity of learning low-loss predictors
representation learning representation evaluation unsupervised learning self-supervised learning
We consider the problem of evaluating representations of data for use in solving a downstream task. We propose to measure the quality of a representation by the complexity of learning a predictor on top of the representation that achieves low loss on a task of interest. To this end, we introduce two measures: surplus description length (SDL) and $\varepsilon$ sample complexity ($\varepsilon$SC). To compare our methods to prior work, we also present a framework based on plotting the validation loss versus dataset size (the "loss-data" curve). Existing measures, such as mutual information and minimum description length, correspond to slices and integrals along the data-axis of the loss-data curve, while ours correspond to slices and integrals along the loss-axis. This analysis shows that prior methods measure properties of an evaluation dataset of a specified size, whereas our methods measure properties of a predictor with a specified loss. We conclude with experiments on real data to compare the behavior of these methods over datasets of varying size.
|
false |
CaLFADS: latent factor analysis of dynamical systems in calcium imaging data
latent variable modelling lfads neuroscience variational autoencoders dynamical systems calcium imaging neural data analysis
Dynamic latent variable modelling has been a hugely powerful tool in understanding how spiking activity in populations of neurons can perform computations necessary for adaptive behaviour. The success of such approaches has been enabled by the ability to construct models derived with the characterization of spiking activity as point-processes since spiking dynamics occur on a much faster time-scale than the computational dynamics being inferred. Other experimental techniques, such as calcium imaging, pose a problem for latent variable modelling of computational dynamics, since the time-scales of calcium dynamics and computational dynamics overlap. As such, the success of dynamic latent variable modelling in calcium imaging data rests on being able to disentangle the contribution of these two sources of variation. Here we extend recent advances using variational autoencoders to analyze neural data, by incorporating a ladder architecture that can infer a hierarchy of dynamical systems. Using built-in inductive biases for calcium dynamics, we can capture calcium flux as well as underlying dynamics of neural computation. First, we demonstrate with synthetic calcium data that we can correctly infer an underlying Lorenz attractor at the same time as calcium dynamics. Next, we show that we can infer appropriate rotational dynamics in spiking data from macaque motor cortex after it has been converted into calcium fluorescence data via a calcium dynamics model. Finally, we show that our method applied to real calcium imaging data from primary visual cortex in mice allows us to infer latent factors that carry salient sensory information about unexpected stimuli. These results demonstrate that variational ladder autoencoders are a promising approach for inferring hierarchical dynamics in experimental settings where the measured variable has its own slow dynamics, such as calcium imaging data, thereby providing the neuroscience community with a new analysis tool for a wider array of data modalities.
|
false |
ItNet: iterative neural networks for fast and efficient anytime prediction
efficient deep neural network semantic segmentation parameter sharing anytime prediction tiny network graph massively parallel hardware systems recurrent convolutional network
Deep neural networks have usually to be compressed and accelerated for their usage in low-power, e.g. mobile, devices. Common requirements are high accuracy, high throughput, low latency, and a small memory footprint. A good trade-off between accuracy and latency has been shown by networks comprising multiple intermediate outputs. In this study, we introduce a multi-output network that has a tiny memory footprint in terms of its computational graph, which allows its execution on novel, massively-parallel hardware accelerators designed for extremely high throughput. To this end, the graph is designed to contain loops by iteratively executing a single network building block. These so-called iterative neural networks enable stateof-the-art results for semantic segmentation on the CamVid and Cityscapes datasets that are especially demanding in terms of computational resources. In ablation studies, the improvement of network training by intermediate network outputs as well as the trade-off between weight sharing over iterations and the network size are investigated.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.