output
bool 2
classes | input
stringlengths 345
2.91k
|
|---|---|
true |
A Block Minifloat Representation for Training Deep Neural Networks
deep neural networks representations block minifloat representation operations accuracy dnn high efficiency difficult native
Training Deep Neural Networks (DNN) with high efficiency can be difficult to achieve with native floating-point representations and commercially available hardware. Specialized arithmetic with custom acceleration offers perhaps the most promising alternative. Ongoing research is trending towards narrow floating-point representations, called minifloats, that pack more operations for a given silicon area and consume less power. In this paper, we introduce Block Minifloat (BM), a new spectrum of minifloat formats capable of training DNNs end-to-end with only 4-8 bit weight, activation and gradient tensors. While standard floating-point representations have two degrees of freedom, via the exponent and mantissa, BM exposes the exponent bias as an additional field for optimization. Crucially, this enables training with fewer exponent bits, yielding dense integer-like hardware for fused multiply-add (FMA) operations. For ResNet trained on ImageNet, 6-bit BM achieves almost no degradation in floating-point accuracy with FMA units that are $4.1\times(23.9\times)$ smaller and consume $2.3\times(16.1\times)$ less energy than FP8 (FP32). Furthermore, our 8-bit BM format matches floating-point accuracy while delivering a higher computational density and faster expected training times.
|
true |
Data-Driven Higher Order Differential Equations Inspired Graph Neural Networks
Graph Neural Networks Differential Equations
A recent innovation in Graph Neural Networks (GNNs) is the family of Differential Equation-Inspired Graph Neural Networks (DE-GNNs), which leverage principles from continuous dynamical systems to model information flow on graphs with built-in properties. However, existing DE-GNNs rely on first or second-order temporal orders. In this paper, we propose a neural extension to those pre-defined temporal dependencies. We show that our model, called TDE-GNN, can capture a wide range of temporal dynamics that go beyond typical first or second-order methods, and provide use cases where existing temporal models are challenged. We demonstrate the benefit of learning the temporal dependencies using our method rather than using pre-defined temporal dynamics on several graph benchmarks.
|
false |
Generative Fairness Teaching
fairness student teacher model counterfactual generative model
Increasing evidences has shown that data biases towards sensitive features such as gender or race are often inherited or even amplified by machine learning models. Recent advancements in fairness mitigate such biases by adjusting the predictions across sensitive groups during the training. Such a correction, however, can only take advantage of samples in a fixed dataset, which usually has limited amount of samples for the minority groups. We propose a generative fairness teaching framework that provides a model with not only real samples but also synthesized samples to compensate the data biases during training. We employ such a teaching strategy by implementing a Generative Fairness Teacher (GFT) that dynamically adjust the proportion of training data for a biased student model. Experimental results indicated that our teacher model is capable of guiding a wide range of biased models by improving the fairness and performance trade-offs significantly.
|
false |
Neuron Activation Analysis for Multi-Joint Robot Reinforcement Learning
Reinforcement Learning Machine Learning Robot Motion Learning DQN Robot Manipulator Target Reaching Network Pruning
Recent experiments indicate that pre-training of end-to-end Reinforcement Learning neural networks on general tasks can speed up the training process for specific robotic applications. However, it remains open if these networks form general feature extractors and a hierarchical organization that are reused as apparent e.g. in Convolutional Neural Networks. In this paper we analyze the intrinsic neuron activation in networks trained for target reaching of robot manipulators with increasing joint number in a vertical plane. We analyze the individual neuron activity distribution in the network, introduce a pruning algorithm to reduce network size keeping the performance, and with these dense network representations we spot correlations of neuron activity patterns among networks trained for robot manipulators with different joint number. We show that the input and output network layers have more distinct neuron activation in contrast to inner layers. Our pruning algorithm reduces the network size significantly, increases the distance of neuron activation while keeping a high performance in training and evaluation. Our results demonstrate that neuron activity can be mapped among networks trained for robots with different complexity. Hereby, robots with small joint difference show higher layer-wise projection accuracy whereas more different robots mostly show projections to the first layer.
|
true |
Applications of Fourier Neural Operators in the Ifmif-Dones Accelerator
Fourier Neural Operators IFMIF-DONES accelerator Deep Learning Deep Reinforcement Learning Surrogate Models Parameter Optimization Linear Accelerator Beam Control Inverse Problem Accelerated Simulation Nuclear Fusion Neutron Irradiation Partial Differential Equations Sequential Decision Making PyTorch Stochastic Gradient Descent Markov Decision Processes Bellman Equations Policy Function Magnetic Control Quadrupoles OPAL Simulator Gymnasium Framework Proximal Policy Optimization Magnetic Forces Particle Distribution Simulation Speed-up Loss Function AI for Science
In this work, Fourier Neural Operators are employed to improve control and optimization of an experimental module of the IFMIF-DONES linear accelerator, otherwise hindered by its simulations high complexity. The models are trained to predict beam envelopes along the lattice’s longitudinal axis, considering variations in quadrupole strengths and particle injections. They serve three purposes: enabling fast inference of beam envelopes, creating an environment for training a Deep Reinforcement Learning agent responsible for shaping the beam, and developing an optimizer for identifying optimal accelerator parameters. The resulting models offer significantly faster predictions (up to 3 orders of magnitude) com- pared to traditional simulators, with maximum percentage errors below 2 %. This accelerated simulation capability makes it feasible to train control agents, since the time per step taken is reduced from 3s to 4 × 10e−3s. Additionally, Stochastic Gradient Descent was applied to optimize one of the models itself, determining the best parameters for a given target and thus solving the inverse problem within seconds. These results demonstrate the synergy and potential of these Deep Learning models, offering promising pathways to advance control and optimization strategies in the IFMIF-DONES accelerator and in other complex scientific facilities.
|
true |
Global Convergence of Three-layer Neural Networks in the Mean Field Regime
deep learning theory
In the mean field regime, neural networks are appropriately scaled so that as the width tends to infinity, the learning dynamics tends to a nonlinear and nontrivial dynamical limit, known as the mean field limit. This lends a way to study large-width neural networks via analyzing the mean field limit. Recent works have successfully applied such analysis to two-layer networks and provided global convergence guarantees. The extension to multilayer ones however has been a highly challenging puzzle, and little is known about the optimization efficiency in the mean field regime when there are more than two layers.
In this work, we prove a global convergence result for unregularized feedforward three-layer networks in the mean field regime. We first develop a rigorous framework to establish the mean field limit of three-layer networks under stochastic gradient descent training. To that end, we propose the idea of a neuronal embedding, which comprises of a fixed probability space that encapsulates neural networks of arbitrary sizes. The identified mean field limit is then used to prove a global convergence guarantee under suitable regularity and convergence mode assumptions, which – unlike previous works on two-layer networks – does not rely critically on convexity. Underlying the result is a universal approximation property, natural of neural networks, which importantly is shown to hold at any finite training time (not necessarily at convergence) via an algebraic topology argument.
|
false |
Physics-Aware Difference Graph Networks for Sparsely-Observed Dynamics
PA-DGN PDE Graph Networks Sparse Data RGN
Sparsely available data points cause numerical error on finite differences which hinders us from modeling the dynamics of physical systems. The discretization error becomes even larger when the sparse data are irregularly distributed or defined on an unstructured grid, making it hard to build deep learning models to handle physics-governing observations on the unstructured grid. In this paper, we propose a novel architecture, Physics-aware Difference Graph Networks (PA-DGN), which exploits neighboring information to learn finite differences inspired by physics equations. PA-DGN leverages data-driven end-to-end learning to discoverunderlying dynamical relations between the spatial and temporal differences in given sequential observations. We demonstrate the superiority of PA-DGN in theapproximation of directional derivatives and the prediction of graph signals on the synthetic data and the real-world climate observations from weather stations.
|
false |
Efficient Reinforcement Learning in Resource Allocation Problems Through Permutation Invariant Multi-task Learning
learning efficient reinforcement learning resource allocation problems permutation invariant main challenges reinforcement learning limited training samples certain settings available data form
One of the main challenges in real-world reinforcement learning is to learn successfully from limited training samples. We show that in certain settings, the available data can be dramatically increased through a form of multi-task learning, by exploiting an invariance property in the tasks. We provide a theoretical performance bound for the gain in sample efficiency under this setting. This motivates a new approach to multi-task learning, which involves the design of an appropriate neural network architecture and a prioritized task-sampling strategy. We demonstrate empirically the effectiveness of the proposed approach on two real-world sequential resource allocation tasks where this invariance property occurs: financial portfolio optimization and meta federated learning.
|
true |
Disentangling 3D Prototypical Networks for Few-Shot Concept Learning
Disentanglement Few Shot Learning 3D Vision VQA
We present neural architectures that disentangle RGB-D images into objects’ shapes and styles and a map of the background scene, and explore their applications for few-shot 3D object detection and few-shot concept classification. Our networks incorporate architectural biases that reflect the image formation process, 3D geometry of the world scene, and shape-style interplay. They are trained end-to-end self-supervised by predicting views in static scenes, alongside a small number of 3D object boxes. Objects and scenes are represented in terms of 3D feature grids in the bottleneck of the network. We show the proposed 3D neural representations are compositional: they can generate novel 3D scene feature maps by mixing object shapes and styles, resizing and adding the resulting object 3D feature maps over background scene feature maps. We show object detectors trained on hallucinated 3D neural scenes generalize better to novel environments. We show classifiers for object categories, color, materials, and spatial relationships trained over the disentangled 3D feature sub-spaces generalize better with dramatically fewer exemplars over the current state-of-the-art, and enable a visual question answering system that uses them as its modules to generalize one-shot to novel objects in the scene.
|
false |
Gradient-based tuning of Hamiltonian Monte Carlo hyperparameters
Hamiltonian Monte Carlo HMC MCMC Variational Inference
Hamiltonian Monte Carlo (HMC) is one of the most successful sampling methods in machine learning. However, its performance is significantly affected by the choice of hyperparameter values, which require careful tuning. Existing approaches for automating this task either optimise a proxy for mixing speed or consider the HMC chain as an implicit variational distribution and optimize a tractable lower bound that is too loose to be useful in practice. Instead, we propose to optimize an objective that quantifies directly the speed of convergence to the target distribution. Our objective can be easily optimized using stochastic gradient descent. We evaluate our proposed method and compare to baselines on a variety of problems including synthetic 2D distributions, the posteriors of variational autoencoders and the Boltzmann distribution for molecular configurations of a 22 atom molecule. We find our method is competitive with or improves upon alternative baselines on all problems we consider.
|
true |
Disentangling Linguistic Features with Dimension-Wise Analysis of Vector Embeddings
interpretability embeddings BERT GPT LLM natural language processing linguistics disentanglement
Understanding the inner workings of neural embeddings, particularly in models such as BERT, remains a challenge because of their high-dimensional and opaque nature. This paper proposes a framework for uncovering the specific dimensions of vector embeddings that encode distinct linguistic properties (LPs). We introduce the Linguistically Distinct Sentence Pairs (LDSP-10) dataset, which isolates ten key linguistic features such as synonymy, negation, tense, and quantity. Using this dataset, we analyze BERT embeddings with various statistical methods, including the Wilcoxon signed-rank test, mutual information, and recursive feature elimination, to identify the most influential dimensions for each LP. We introduce a new metric, the Embedding Dimension Importance (EDI) score, which quantifies the relevance of each embedding dimension to a LP. Our findings show that certain properties, such as negation and polarity, are robustly encoded in specific dimensions, while others, like synonymy, exhibit more complex patterns. This study provides insights into the interpretability of embeddings, which can guide the development of more transparent and optimized language models, with implications for model bias mitigation and the responsible deployment of AI systems.
|
false |
Deterministic Policy Imitation Gradient Algorithm
Imitation Learning
The goal of imitation learning (IL) is to enable a learner to imitate an expert’s behavior given the expert’s demonstrations. Recently, generative adversarial imitation learning (GAIL) has successfully achieved it even on complex continuous control tasks. However, GAIL requires a huge number of interactions with environment during training. We believe that IL algorithm could be more applicable to the real-world environments if the number of interactions could be reduced. To this end, we propose a model free, off-policy IL algorithm for continuous control. The keys of our algorithm are two folds: 1) adopting deterministic policy that allows us to derive a novel type of policy gradient which we call deterministic policy imitation gradient (DPIG), 2) introducing a function which we call state screening function (SSF) to avoid noisy policy updates with states that are not typical of those appeared on the expert’s demonstrations. Experimental results show that our algorithm can achieve the goal of IL with at least tens of times less interactions than GAIL on a variety of continuous control tasks.
|
true |
Inductive Biases for Relational Tasks
Out-of-distribution generalization inductive biases neuro-inspired AI relational reasoning
Current deep learning approaches have shown good in-distribution performance but struggle in out-of-distribution settings. This is especially true in the case of tasks involving abstract relations like recognizing rules in sequences, as required in many intelligence tests. In contrast, our brains are remarkably flexible at such tasks, an attribute that is likely linked to anatomical constraints on computations. Inspired by this, recent work has explored how enforcing that relational representations remain distinct from sensory representations can help artificial systems. Building on this work, we further explore and formalize the advantages afforded by ``partitioned'' representations of relations and sensory details. We investigate inductive biases that ensure abstract relations are learned and represented distinctly from sensory data across several neural network architectures and show that they outperform existing architectures on out-of-distribution generalization for various relational tasks. These results show that partitioning relational representations from other information streams may be a simple way to augment existing network architectures' robustness when performing relational computations.
|
true |
Learning to reason about and to act on physical cascading events
events reason physical challenging cascades cascade agent dynamic environments fundamental problem
Reasoning and interacting with dynamic environments is a fundamental problem in AI, but it becomes extremely challenging when actions can trigger cascades of cross-dependant events.
We introduce a new learning setup called Cascade where an agent is shown a video of a simulated physical dynamic scene, and is asked to intervene and trigger a cascade of events, such that the system reaches a "counterfactual" goal. For instance, the agent may be asked to “Make the blue ball hit the red one, by pushing the green ball”. The problem is very challenging because agent interventions are from a continuous space, and cascades of events make the dynamics highly non-linear.
We combine semantic tree search with an event-driven forward model and devise an algorithm that learns to search in semantic trees in continuous spaces. We demonstrate that our approach learns to effectively follow instructions to intervene in previously unseen complex scenes. Interestingly, it can use the observed cascade of events to reason about alternative counterfactual outcomes.
|
false |
Decomposing Mutual Information for Representation Learning
Mutual Information Self-supervised learning
Many self-supervised representation learning methods maximize mutual information (MI) across views. In this paper, we transform each view into a set of subviews and then decompose the original MI bound into a sum of bounds involving conditional MI between the subviews. E.g.,~given two views $x$ and $y$ of the same input example, we can split $x$ into two subviews, $x^{\prime}$ and $x^{\prime\prime}$, which depend only on $x$ but are otherwise unconstrained. The following holds: $I(x; y) \geq I(x^{\prime\prime}; y) + I(x^{\prime}; y | x^{\prime\prime})$, due to the chain rule and information processing inequality. By maximizing both terms in the decomposition, our approach explicitly rewards the encoder for any information about $y$ which it extracts from $x^{\prime\prime}$, and for information about $y$ extracted from $x^{\prime}$ in excess of the information from $x^{\prime\prime}$. We provide a novel contrastive lower-bound on conditional MI, that relies on sampling contrast sets from $p(y|x^{\prime\prime})$. By decomposing the original MI into a sum of increasingly challenging MI bounds between sets of increasingly informed views, our representations can capture more of the total information shared between the original views. We empirically test the method in a vision domain and for dialogue generation.
|
false |
Effective Regularization Through Loss-Function Metalearning
regularization loss loss function metalearning meta-learning optimization theory robustness adversarial attacks
Loss-function metalearning can be used to discover novel, customized loss functions for deep neural networks, resulting in improved performance, faster training, and improved data utilization. A likely explanation is that such functions discourage overfitting, leading to effective regularization. This paper theoretically demonstrates that this is indeed the case: decomposition of learning rules makes it possible to characterize the training dynamics and show that loss functions evolved through TaylorGLO regularize both in the beginning and end of learning, and maintain an invariant in between. The invariant can be utilized to make the metalearning process more efficient in practice, and the regularization can train networks that are robust against adversarial attacks. Loss-function optimization can thus be seen as a well-founded new aspect of metalearning in neural networks.
|
true |
Antipodal Pairing and Mechanistic Signals in Dense SAE Latents
interpretability mechanistic interpretability SAE sparse autoencoder
Sparse autoencoders (SAEs) are designed to extract interpretable features from language models, yet they often yield frequently activating latents that remain difficult to interpret. It is still an open question whether these \textit{dense} latents are an undesired training artifact or whether they represent fundamentally dense signals in the model's activations. Our study provides evidence for the latter explanation:
dense latents capture fundamental signals which (1) align with principal directions of variance in the model's residual stream and (2) reconstruct a subspace of the unembedding matrix that was linked by previous work to internal model computation.
Furthermore, we show that these latents typically emerge as nearly antipodal pairs that collaboratively reconstruct specific residual stream directions. These findings reveal a mechanistic role for dense latents in language model behavior and suggest avenues for refining SAE training strategies.
|
false |
Deep Positive Unlabeled Learning with a Sequential Bias
sequential bias deepspu datasets methods positive unlabeled learning many domains video stream analytics human activity recognition available
For many domains, from video stream analytics to human activity recognition, only weakly-labeled datasets are available.
Worse yet, the given labels are often assigned sequentially, resulting in sequential bias. Current Positive Unlabeled (PU) classifiers, a state-of-the-art family of robust semi-supervised methods, are ineffective under sequential bias. In this work, we propose DeepSPU, the first method to address this sequential bias problem. DeepSPU tackles the two interdependent subproblems of learning both the latent labeling process and the true class likelihoods within one architecture. We achieve this by developing a novel iterative learning strategy aided by theoretically-justified cost terms to avoid collapsing into a naive classifier. Our experimental studies demonstrate that DeepSPU outperforms state-of-the-art methods by over 10% on diverse real-world datasets.
|
true |
FairFil: Contrastive Neural Debiasing Method for Pretrained Text Encoders
Fairness Contrastive Learning Mutual Information Pretrained Text Encoders
Pretrained text encoders, such as BERT, have been applied increasingly in various natural language processing (NLP) tasks, and have recently demonstrated significant performance gains. However, recent studies have demonstrated the existence of social bias in these pretrained NLP models. Although prior works have made progress on word-level debiasing, improved sentence-level fairness of pretrained encoders still lacks exploration. In this paper, we proposed the first neural debiasing method for a pretrained sentence encoder, which transforms the pretrained encoder outputs into debiased representations via a fair filter (FairFil) network. To learn the FairFil, we introduce a contrastive learning framework that not only minimizes the correlation between filtered embeddings and bias words but also preserves rich semantic information of the original sentences. On real-world datasets, our FairFil effectively reduces the bias degree of pretrained text encoders, while continuously showing desirable performance on downstream tasks. Moreover, our post hoc method does not require any retraining of the text encoders, further enlarging FairFil's application space.
|
true |
Complex Query Answering with Neural Link Predictors
neural link prediction complex query answering
Neural link predictors are immensely useful for identifying missing edges in large scale Knowledge Graphs. However, it is still not clear how to use these models for answering more complex queries that arise in a number of domains, such as queries using logical conjunctions ($\land$), disjunctions ($\lor$) and existential quantifiers ($\exists$), while accounting for missing edges. In this work, we propose a framework for efficiently answering complex queries on incomplete Knowledge Graphs. We translate each query into an end-to-end differentiable objective, where the truth value of each atom is computed by a pre-trained neural link predictor. We then analyse two solutions to the optimisation problem, including gradient-based and combinatorial search. In our experiments, the proposed approach produces more accurate results than state-of-the-art methods --- black-box neural models trained on millions of generated queries --- without the need of training on a large and diverse set of complex queries. Using orders of magnitude less training data, we obtain relative improvements ranging from 8% up to 40% in Hits@3 across different knowledge graphs containing factual information. Finally, we demonstrate that it is possible to explain the outcome of our model in terms of the intermediate solutions identified for each of the complex query atoms. All our source code and datasets are available online, at https://github.com/uclnlp/cqd.
|
false |
RetCL: A Selection-based Approach for Retrosynthesis via Contrastive Learning
molecule retrosynthesis contrastive learning graph representation learning
Retrosynthesis, of which the goal is to find a set of reactants for synthesizing a target product, is an emerging research area of deep learning. While the existing approaches have shown promising results, they currently lack the ability to consider availability (e.g., stability or purchasability) of the reactants or generalize to unseen reaction templates (i.e., chemical reaction rules).
In this paper, we propose a new approach that mitigates the issues by reformulating retrosynthesis into a selection problem of reactants from a candidate set of commercially available molecules. To this end, we design an efficient reactant selection framework, named RetCL (retrosynthesis via contrastive learning), for enumerating all of the candidate molecules based on selection scores computed by graph neural networks. For learning the score functions, we also propose a novel contrastive training scheme with hard negative mining. Extensive experiments demonstrate the benefits of the proposed selection-based approach. For example, when all 671k reactants in the USPTO database are given as candidates, our RetCL achieves top-1 exact match accuracy of 71.3% for the USPTO-50k benchmark, while a recent transformer-based approach achieves 59.6%. We also demonstrate that RetCL generalizes well to unseen templates in various settings in contrast to template-based approaches. The code will be released.
|
false |
Learning Blood Oxygen from Respiration Signals
healthcare medical application
Monitoring blood oxygen is critical in a variety of medical conditions. For almost a century, pulse oximetry has been the only non-invasive method for measuring blood oxygen. While highly useful, pulse oximetry has important limitations. It requires wearable sensors, which can be cumbersome for older patients. It is also known to be biased when used for dark-skinned subjects. In this paper, we demonstrate, for the first time, the feasibility of predicting oxygen saturation from breathing. By eliminating the dependency on oximetry, we eliminate bias against skin color. Further, since breathing can be monitored without body contact by analyzing the radio signal in the environment, we show that oxygen too can be monitored without any wearable devices. We introduce a new approach for leveraging auxiliary variables via a switcher-based multi-headed neural network model. Empirical results show that our model achieves good accuracy on multiple medical datasets.
|
true |
Seeing is Not Necessarily Believing: Limitations of BigGANs for Data Augmentation
GANs data augmentation BigGAN
Recent advances in Generative Adversarial Networks (GANs) – in architectural design, training strategies, and empirical tricks – have led nearly photorealistic samples on large-scale datasets such as ImageNet. In fact, for one model in particular, BigGAN, metrics such as Inception Score or Frechet Inception Distance nearly match those of the dataset, suggesting that these models are close to match-ing the distribution of the training set. Given the quality of these models, it is worth understanding to what extent these samples can be used for data augmentation, a task expressed as a long-term goal of the GAN research project. To that end, we train ResNet-50 classifiers using either purely BigGAN images or mixtures of ImageNet and BigGAN images, and test on the ImageNet validation set.Our preliminary results suggest both a measured view of state-of-the-art GAN quality and highlight limitations of current metrics. Using only BigGAN images, we find that Top-1 and Top-5 error increased by 120% and 384%, respectively, and furthermore, adding more BigGAN data to the ImageNet training set at best only marginally improves classifier performance. Finally, we find that neither Inception Score, nor FID, nor combinations thereof are predictive of classification accuracy. These results suggest that as GANs are beginning to be deployed in downstream tasks, we should create metrics that better measure downstream task performance. We propose classification performance as one such metric that, in addition to assessing per-class sample quality, is more suited to such downstream tasks.
|
false |
A Unified Spectral Sparsification Framework for Directed Graphs
Spectral Graph Theory Spectral Sparsification Directed Graphs Laplacian Solver PageRank Vectors
Recent spectral graph sparsification research allows constructing nearly-linear-sized subgraphs that can well preserve the spectral (structural) properties of the original graph, such as the first few eigenvalues and eigenvectors of the graph Laplacian, leading to the development of a variety of nearly-linear time numerical and graph algorithms. However, there is not a unified approach that allows for truly scalable spectral sparsification of both directed and undirected graphs. For the first time, we prove the existence of linear-sized spectral sparsifiers for general directed
graphs and introduce a practically-efficient and unified spectral graph sparsification approach that allows sparsifying real-world, large-scale directed and undirected graphs with guaranteed preservation of the original graph spectra. By exploiting a highly-scalable (nearly-linear complexity) spectral matrix perturbation analysis framework for constructing nearly-linear sized (directed) subgraphs, it enables us to well preserve the key eigenvalues and eigenvectors of the original (directed) graph
Laplacians. The proposed method has been validated using various kinds of directed graphs obtained from public domain sparse matrix collections, showing promising results for solving directed graph Laplacians, spectral embedding, and partitioning of general directed graphs, as well as approximately computing (personalized) PageRank vectors.
|
true |
Estimating Lipschitz constants of monotone deep equilibrium models
deep equilibrium models Lipschitz constants
Several methods have been proposed in recent years to provide bounds on the Lipschitz constants of deep networks, which can be used to provide robustness guarantees, generalization bounds, and characterize the smoothness of decision boundaries. However, existing bounds get substantially weaker with increasing depth of the network, which makes it unclear how to apply such bounds to recently proposed models such as the deep equilibrium (DEQ) model, which can be viewed as representing an infinitely-deep network. In this paper, we show that monotone DEQs, a recently-proposed subclass of DEQs, have Lipschitz constants that can be bounded as a simple function of the strong monotonicity parameter of the network. We derive simple-yet-tight bounds on both the input-output mapping and the weight-output mapping defined by these networks, and demonstrate that they are small relative to those for comparable standard DNNs. We show that one can use these bounds to design monotone DEQ models, even with e.g. multi-scale convolutional structure, that still have constraints on the Lipschitz constant. We also highlight how to use these bounds to develop PAC-Bayes generalization bounds that do not depend on any depth of the network, and which avoid the exponential depth-dependence of comparable DNN bounds.
|
true |
Sifting the Signal from the Noise
signaling game reinforcement learning attention information
Signaling games are useful for understanding how language emerges. In the standard models the dynamics in some sense already knows what the signals are, even if they do not yet have meaning. In this paper we develop a simple model we call an attention game in which agents have to learn which feature in their environment is the signal. We demonstrate that simple reinforcement learning agents can still learn to coordinate in contexts in which (i) the agents do not already know what the signal is and (ii) the other features in the agents’ environment are uncorrelated with the signal. Furthermore, we show that, in the cases in which other features are correlated with the signal, there is a surprising trade-off between learning to pay attention to the signal and success in action. We show that the mutual information between a signal and a feature plays a key role in governing the accuracy and attention of the agent.
|
false |
A Gradient-based Kernel Approach for Efficient Network Architecture Search
NAS
It is widely accepted that vanishing and exploding gradient values are the main reason behind the difficulty of deep network training.
In this work, we take a further step to understand the optimization of deep networks and find that both gradient correlations and gradient values have strong impacts on model training.
Inspired by our new finding, we explore a simple yet effective network architecture search (NAS) approach that leverages gradient correlation and gradient values to find well-performing architectures. To be specific, we first formulate these two terms into a unified gradient-based kernel and then select architectures with the largest kernels at initialization as the final networks.
The new approach replaces the expensive ``train-then-test'' evaluation paradigm with a new lightweight function according to the gradient-based kernel at initialization.
Experiments show that our approach achieves competitive results with orders of magnitude faster than ``train-then-test'' paradigms on image classification tasks. Furthermore, the extremely low search cost enables its wide applications. It also obtains performance improvements on two text classification tasks.
|
false |
Meta Auxiliary Labels with Constituent-based Transformer for Aspect-based Sentiment Analysis
Natural Language Processing Sentiment Analysis
Aspect based sentiment analysis (ABSA) is a challenging natural language processing task that could benefit from syntactic information. Previous work exploit dependency parses to improve performance on the task, but this requires the existence of good dependency parsers. In this paper, we build a constituent-based transformer for ABSA that can induce constituents without constituent parsers. We also apply meta auxiliary learning to generate labels on edges between tokens, supervised by the objective of the ABSA task. Without input from dependency parsers, our models outperform previous work on three Twitter data sets and match previous work closely on two review data sets.
|
false |
Latent Space Semi-Supervised Time Series Data Clustering
Semi-supervised clustering clustering deep learning autoencoder
Time series data is abundantly available in the real world, but there is a distinct lack of large, labeled datasets available for many types of learning tasks. Semi-supervised models, which can leverage small amounts of expert-labeled data along with a larger unlabeled dataset, have been shown to improve performance over unsupervised learning models. Existing semi-supervised time series clustering algorithms suffer from lack of scalability as they are limited to perform learning operations within the original data space. We propose an autoencoder-based semi-supervised learning model along with multiple semi-supervised objective functions which can be used to improve the quality of the autoencoder’s learned latent space via the addition of a small number of labeled examples. Experiments on a variety of datasets show that our methods can usually improve k-Means clustering performance. Our methods achieve a maximum average ARI of 0.897, a 140% increase over an unsupervised CAE model. Our methods also achieve a maximum improvement of 44% over a semi-supervised model.
|
true |
LogicInference: A new Datasaet for Teaching Logical Inference to seq2seq Models
dataset logical inference reasoning transformers
Machine learning models such as Transformers or LSTMs struggle with tasks that are compositional in nature such as those involving reasoning/inference. Although many datasets exist to evaluate compositional generalization, when it comes to evaluating inference abilities, options are more limited. This paper presents LogicInference, a new dataset to evaluate the ability of models to perform logical inference. The dataset focuses on inference using propositional logic and a small subset of first-order logic, represented both in semi-formal logical notation, as well as in natural language. We also report initial results using a collection of machine learning models to establish an initial baseline in this dataset.
|
true |
Expressive Power of Invariant and Equivariant Graph Neural Networks
Graph Neural Network Universality Approximation
Various classes of Graph Neural Networks (GNN) have been proposed and shown to be successful in a wide range of applications with graph structured data. In this paper, we propose a theoretical framework able to compare the expressive power of these GNN architectures. The current universality theorems only apply to intractable classes of GNNs. Here, we prove the first approximation guarantees for practical GNNs, paving the way for a better understanding of their generalization. Our theoretical results are proved for invariant GNNs computing a graph embedding (permutation of the nodes of the input graph does not affect the output) and equivariant GNNs computing an embedding of the nodes (permutation of the input permutes the output). We show that Folklore Graph Neural Networks (FGNN), which are tensor based GNNs augmented with matrix multiplication are the most expressive architectures proposed so far for a given tensor order. We illustrate our results on the Quadratic Assignment Problem (a NP-Hard combinatorial problem) by showing that FGNNs are able to learn how to solve the problem, leading to much better average performances than existing algorithms (based on spectral, SDP or other GNNs architectures). On a practical side, we also implement masked tensors to handle batches of graphs of varying sizes.
|
true |
DUAL SPACE LEARNING WITH VARIATIONAL AUTOENCODERS
variational autoencoder image transfer multi-class dual space
This paper proposes a dual variational autoencoder (DualVAE), a framework for generating images corresponding to multiclass labels. Recent research on conditional generative models, such as the Conditional VAE, exhibit image transfer by changing labels. However, when the dimension of multiclass labels is large, these models cannot change images corresponding to labels, because learning multiple distributions of the corresponding class is necessary to transfer an image. This leads to the lack of training data. Therefore, instead of conditioning with labels, we condition with latent vectors that include label information. DualVAE divides one distribution of the latent space by linear decision boundaries using labels. Consequently, DualVAE can easily transfer an image by moving a latent vector toward a decision boundary and is robust to the missing values of multiclass labels. To evaluate our proposed method, we introduce a conditional inception score (CIS) for measuring how much an image changes to the target class. We evaluate the images transferred by DualVAE using the CIS in CelebA datasets and demonstrate state-of-the-art performance in a multiclass setting.
|
false |
Trust, but verify: model-based exploration in sparse reward environments
reinforcement learning model-based exploration on-line planning imperfect environment model
We propose $\textit{trust-but-verify}$ (TBV) mechanism, a new method which uses model uncertainty estimates to guide exploration. The mechanism augments graph search planning algorithms by the capacity to deal with learned model's imperfections. We identify certain type of frequent model errors, which we dub $\textit{false loops}$, and which are particularly dangerous for graph search algorithms in discrete environments. These errors impose falsely pessimistic expectations and thus hinder exploration. We confirm this experimentally and show that TBV can effectively alleviate them. TBV combined with MCTS or Best First Search forms an effective model-based reinforcement learning solution, which is able to robustly solve sparse reward problems.
|
false |
Function Contrastive Learning of Transferable Representations
Representations Learning Few-shot Learning Contrastive Learning
Few-shot-learning seeks to find models that are capable of fast-adaptation to novel tasks which are not encountered during training. Unlike typical few-shot learning algorithms, we propose a contrastive learning method which is not trained to solve a set of tasks, but rather attempts to find a good representation of the underlying data-generating processes (\emph{functions}). This allows for finding representations which are useful for an entire series of tasks sharing the same function. In particular, our training scheme is driven by the self-supervision signal indicating whether two sets of samples stem from the same underlying function. Our experiments on a number of synthetic and real-world datasets show that the representations we obtain can outperform strong baselines in terms of downstream performance and noise robustness, even when these baselines are trained in an end-to-end manner.
|
true |
Conformalized Physics-Informed Neural Networks
physics-informed neural networks conformal prediction uncertainty quantification
Physics-informed neural networks (PINNs) are an influential method of solving differential equations and estimating their parameters given data. However, since they make use of neural networks, they provide only a point estimate of differential equation parameters, as well as the solution at any given point, without any measure of uncertainty. Ensemble and Bayesian methods have been previously applied to quantify the uncertainty of PINNs, but these methods may require making strong assumptions on the data-generating process, and can be computationally expensive. Here, we introduce Conformalized PINNs (C-PINNs) that, without making any additional assumptions, utilize the framework of conformal prediction to quantify the uncertainty of PINNs by providing intervals that have finite-sample, distribution-free statistical validity.
|
true |
Image GANs meet Differentiable Rendering for Inverse Graphics and Interpretable 3D Neural Rendering
Differentiable rendering inverse graphics GANs
Differentiable rendering has paved the way to training neural networks to perform “inverse graphics” tasks such as predicting 3D geometry from monocular photographs. To train high performing models, most of the current approaches rely on multi-view imagery which are not readily available in practice. Recent Generative Adversarial Networks (GANs) that synthesize images, in contrast, seem to acquire 3D knowledge implicitly during training: object viewpoints can be manipulated by simply manipulating the latent codes. However, these latent codes often lack further physical interpretation and thus GANs cannot easily be inverted to perform explicit 3D reasoning. In this paper, we aim to extract and disentangle 3D knowledge learned by generative models by utilizing differentiable renderers. Key to our approach is to exploit GANs as a multi-view data generator to train an inverse graphics network using an off-the-shelf differentiable renderer, and the trained inverse graphics network as a teacher to disentangle the GAN's latent code into interpretable 3D properties. The entire architecture is trained iteratively using cycle consistency losses. We show that our approach significantly outperforms state-of-the-art inverse graphics networks trained on existing datasets, both quantitatively and via user studies. We further showcase the disentangled GAN as a controllable 3D “neural renderer", complementing traditional graphics renderers.
|
true |
Distributional Sliced-Wasserstein and Applications to Generative Modeling
Deep generative models Sliced Wasserstein Optimal Transport
Sliced-Wasserstein distance (SW) and its variant, Max Sliced-Wasserstein distance (Max-SW), have been used widely in the recent years due to their fast computation and scalability even when the probability measures lie in a very high dimensional space. However, SW requires many unnecessary projection samples to approximate its value while Max-SW only uses the most important projection, which ignores the information of other useful directions. In order to account for these weaknesses, we propose a novel distance, named Distributional Sliced-Wasserstein distance (DSW), that finds an optimal distribution over projections that can balance between exploring distinctive projecting directions and the informativeness of projections themselves. We show that the DSW is a generalization of Max-SW, and it can be computed efficiently by searching for the optimal push-forward measure over a set of probability measures over the unit sphere satisfying certain regularizing constraints that favor distinct directions. Finally, we conduct extensive experiments with large-scale datasets to demonstrate the favorable performances of the proposed distances over the previous sliced-based distances in generative modeling applications.
|
true |
Understanding Over-parameterization in Generative Adversarial Networks
GAN Over-parameterization min-max optimization
A broad class of unsupervised deep learning methods such as Generative Adversarial Networks (GANs) involve training of overparameterized models where the number of parameters of the model exceeds a certain threshold. Indeed, most successful GANs used in practice are trained using overparameterized generator and discriminator networks, both in terms of depth and width. A large body of work in supervised learning have shown the importance of model overparameterization in the convergence of the gradient descent (GD) to globally optimal solutions. In contrast, the unsupervised setting and GANs in particular involve non-convex concave mini-max optimization problems that are often trained using Gradient Descent/Ascent (GDA).
The role and benefits of model overparameterization in the convergence of GDA to a global saddle point in non-convex concave problems is far less understood. In this work, we present a comprehensive analysis of the importance of model overparameterization in GANs both theoretically and empirically. We theoretically show that in an overparameterized GAN model with a $1$-layer neural network generator and a linear discriminator, GDA converges to a global saddle point of the underlying non-convex concave min-max problem. To the best of our knowledge, this is the first result for global convergence of GDA in such settings. Our theory is based on a more general result that holds for a broader class of nonlinear generators and discriminators that obey certain assumptions (including deeper generators and random feature discriminators). Our theory utilizes and builds upon a novel connection with the convergence analysis of linear time-varying dynamical systems which may have broader implications for understanding the convergence behavior of GDA for non-convex concave problems involving overparameterized models. We also empirically study the role of model overparameterization in GANs using several large-scale experiments on CIFAR-10 and Celeb-A datasets. Our experiments show that overparameterization improves the quality of generated samples across various model architectures and datasets. Remarkably, we observe that overparameterization leads to faster and more stable convergence behavior of GDA across the board.
|
false |
H-divergence: A Decision-Theoretic Probability Discrepancy Measure
probability divergence two sample test maximum mean discrepancy
Measuring the discrepancy between two probability distributions is a fundamental problem in machine learning and statistics. Based on ideas from decision theory, we investigate a new class of discrepancies that are based on the optimal decision loss. Two probability distributions are different if the optimal decision loss is higher on the mixture distribution than on each individual distribution. We show that this generalizes popular notions of discrepancy measurements such as the Jensen Shannon divergence and the maximum mean discrepancy. We apply our approach to two-sample tests, which evaluates whether two sets of samples come from the same distribution. On various benchmark and real datasets, we demonstrate that tests based on our generalized notion of discrepancy is able to achieve superior test power. We also apply our approach to sample quality evaluation as an alternative to the FID score, and to understanding the effects of climate change on different social and economic activities.
|
true |
LLM AGENTS FOR LITERATURE TO CODE CONVERSION:CASE STUDY OF HEAT EXCHANGER DESIGN
LLM Agents Industrial Design Code Generation
This paper introduces a framework that utilizes large language model (LLM) agents to extract and convert mathematical models from engineering literature into executable code. Autonomous or semi-autonomous conversion of literature into code facilitates downstream tasks such as hypothesis generation, verification, and benchmarking. Focusing on heat exchanger design, our approach efficiently integrates model extraction, code generation, and performance optimization, with minimal human intervention. The system's knowledge base is continuously refined with each new paper, leading to ongoing improvements. Experiments conducted on 115 research articles using the HxAgent approach demonstrate substantial improvements over the previous non-agentic baseline, HxLLM. Although the work is still in progress, the results highlight the potential of agent-driven workflows in advancing scientific discovery.
|
true |
Neural SPH: Improved Neural Modeling of Lagrangian Fluid Dynamics
computational fluid dynamics Lagrangian simulations graph neural networks rollout stability physics modeling
Smoothed particle hydrodynamics (SPH) is omnipresent in modern engineering and scientific disciplines. SPH is a class of Lagrangian schemes that discretize fluid dynamics via finite material points that are tracked through the evolving velocity field. Due to the particle-like nature of the simulation, graph neural networks (GNNs) have emerged as appealing and successful surrogates. However, the practical utility of such GNN-based simulators relies on their ability to faithfully model physics, providing accurate and stable predictions over long time horizons - which is a notoriously hard problem. In this work, we identify particle clustering originating from tensile instabilities as one of the primary pitfalls. Based on these insights, we enhance both training and rollout inference of state-of-the-art GNN-based simulators with varying components from standard SPH solvers, including pressure, viscous, and external force components. All neural SPH-enhanced simulators achieve better performance than the baseline GNNs, often by orders of magnitude, allowing for significantly longer rollouts and significantly better physics modeling. Code available under https://github.com/tumaer/neuralsph.
|
true |
APPA : Agentic Preformulation Pathway Assistant
LLM Preformulation Sciences Drug Development Pharmaceutical Research Agentic Workflows Machine Learning
The design and development of effective drug formulations is a critical process in pharmaceutical research, particularly for small molecule active pharmaceutical ingredients.
This paper introduces a novel agentic preformulation pathway assistant (APPA),
leveraging large language models coupled to experimental databases and a suite of machine learning models to streamline the preformulation process of drug candidates.
APPA successfully integrates domain expertise from scientific publications, databases holding experimental results, and machine learning predictors to reason and propose optimal preformulation strategies based on the current evidence.
This results in case-specific user guidance for the developability assessment of a new drug and directs towards the most promising experimental route,
significantly reducing the time and resources required for the manual collection and analysis of existing evidence.
The approach aims to accelerate the transition of promising compounds from discovery to preclinical and clinical testing.
|
true |
Finding Sparse Autoencoder Representations Of Errors In CoT Prompting
mechanistic interpretability chain-of-thought prompting sparse autoencoders large language models
Current large language models often suffer from subtle, hard-to-detect reasoning
errors in their intermediate chain-of-thought (CoT) steps. These errors include
logical inconsistencies, factual hallucinations, and arithmetic mistakes, which
compromise trust and reliability. While previous research focuses on mechanistic
interpretability for best output, understanding and categorizing internal reasoning
errors remains challenging. The complexity and non-linear nature of these CoT
sequences call for methods to uncover structured patterns hidden within them. As
an initial step, we evaluate Sparse Autoencoder (SAE) activations within neural
networks to investigate how specific neurons contribute to different types of errors.
|
false |
Multimodal Variational Autoencoders for Semi-Supervised Learning: In Defense of Product-of-Experts
variational autoencoder multimodal data product-of-experts semi-supervised learning
Multimodal generative models should be able to learn a meaningful latent representation that enables a coherent joint generation of all modalities (e.g., images and text). Many applications also require the ability to accurately sample modalities conditioned on observations of a subset of the modalities. Often not all modalities may be observed for all training data points, so semi-supervised learning should be possible.
In this study, we evaluate a family of product-of-experts (PoE) based variational autoencoders that have these desired properties. We include a novel PoE based architecture and training procedure. An empirical evaluation shows that the PoE based models can outperform an additive mixture-of-experts (MoE) approach.
Our experiments support the intuition that PoE models are more suited for a conjunctive combination of modalities while MoEs are more suited for a disjunctive fusion.
|
true |
Top of the CLASS: Benchmarking LLM Agents on Real-World Enterprise Tasks
agents llms benchmarks enterprise conversational workflows
Enterprises are increasingly adopting AI agents based on large language models (LLMs) for mission-critical workflows. However, most existing benchmarks use synthetic or consumer-oriented data, and do not holistically evaluate agents on operational concerns beyond accuracy (e.g. cost, security, etc.). To address these gaps we propose CLASSIC, a novel benchmark containing 2,133 real-world user-chatbot conversations and 423 workflows across 7 enterprise domains including IT, HR, banking, and healthcare. We evaluate LLMs across five key metrics -- Cost, Latency, Accuracy, Stability, and Security -- on a multiclass classification task that requires the model to select the proper workflow to trigger in response to a user message. Our dataset of real-world conversations is challenging, with the best LLM achieving an overall accuracy of only 76.1%. Across all five metrics, we find significant variation in performance -- for example, Gemini 1.5 Pro only refuses 78.5% of our jailbreak prompts compared to Claude 3.5 Sonnet's 99.8%, while GPT-4o costs 5.4x more than the most affordable model we evaluate. We hope that our benchmark helps to increase trust in LLM applications by better grounding evaluations in real-world enterprise data. We open source our code and data, and welcome contributions from the community.
|
false |
Learning Twitter User Sentiments on Climate Change with Limited Labeled Data
Climate Change Twitter Data Sentiment Analysis Automated Labelling Cohort Analysis
While it is well-documented that climate change accepters and deniers have become increasingly polarized in the United States over time, there has been no large-scale examination of whether these individuals are prone to changing their opinions as a result of natural external occurrences. On the sub-population of Twitter users, we examine whether climate change sentiment changes in response to five separate natural disasters occurring in the U.S. in 2018. We begin by showing that tweets can be classified with over 75% accuracy as either accepting or denying climate change when using our methodology to compensate for limited labelled data; results are robust across several machine learning models and yield geographic-level results in line with prior research. We then apply RNNs to conduct a cohort-level analysis showing that the 2018 hurricanes yielded a statistically significant increase in average tweet sentiment affirming climate change. However, this effect does not hold for the 2018 blizzard and wildfires studied, implying that Twitter users' opinions on climate change are fairly ingrained on this subset of natural disasters.
|
false |
Length-Adaptive Transformer: Train Once with Length Drop, Use Anytime with Search
transformer efficiency anytime prediction sequence length evolutionary search
Although transformers have achieved impressive accuracies in various tasks in natural language processing, they often come with a prohibitive computational cost, that prevents their use in scenarios with limited computational resources for inference. This need for computational efficiency in inference has been addressed by for instance PoWER-BERT (Goyal et al., 2020) which gradually decreases the length of a sequence as it is passed through layers. These approaches however often assume that the target computational complexity is known in advance at the time of training. This implies that a separate model must be trained for each inference scenario with its distinct computational budget. In this paper, we extend PoWER-BERT to address this issue of inefficiency and redundancy. The proposed extension enables us to train a large-scale transformer, called Length-Adaptive Transformer, once and uses it for various inference scenarios without re-training it. To do so, we train a transformer with LengthDrop, a structural variant of dropout, which stochastically determines the length of a sequence at each layer. We then use a multi-objective evolutionary search to find a length configuration that maximizes the accuracy and minimizes the computational complexity under any given computational budget. Additionally, we significantly extend the applicability of PoWER-BERT beyond sequence-level classification into token-level classification such as span-based question-answering, by introducing the idea of Drop-and-Restore. With Drop-and-Restore, word-vectors are dropped temporarily in intermediate layers and restored at the last layer if necessary. We empirically verify the utility of the proposed approach by demonstrating the superior accuracy-efficiency trade-off under various setups, including SQuAD 1.1, MNLI-m, and SST-2. Upon publication, the code to reproduce our work will be open-sourced.
|
true |
Black-Box Adversarial Attacks on LLM-Based Code Completion
code model black-box attack safety vulnerability large language model jailbreak
Modern code completion engines, powered by large language models (LLMs), assist millions of developers through their strong capabilities to generate functionally correct code. Due to this popularity, it is crucial to investigate the security implications of relying on LLM-based code completion. In this work, we demonstrate that state-of-the-art black-box LLM-based code completion engines can be stealthily biased by adversaries to significantly increase their rate of insecure code generation. We present the first attack, named INSEC, that achieves this goal. INSEC works by injecting an attack string as a short comment in the completion input. The attack string is crafted through a query-based optimization procedure starting from a set of carefully designed initialization schemes. We demonstrate INSEC's broad applicability and effectiveness by evaluating it on various state-of-the-art open-source models and black-box commercial services (e.g., OpenAI API and GitHub Copilot). On a diverse set of security-critical test cases, covering 16 CWEs across 5 programming languages, INSEC increases the rate of generated insecure code by more than 50\%, while maintaining the functional correctness of generated code. INSEC is highly practical -- it requires low resources and costs less than 10 US dollars to develop on commodity hardware. Moreover, we showcase the attack's real-world deployment, by developing an IDE plug-in that stealthily injects INSEC into the GitHub Copilot extension.
|
true |
THE FUNDAMENTAL LIMITS OF LLM UNLEARNING: COMPLEXITY-THEORETIC BARRIERS AND PROVABLY OPTIMAL PROTOCOLS
Machine Unlearning Large Language Models Computational Complexity Differential Privacy GDPR Compliance AI Trustworthiness
Modern machine unlearning techniques for large language models (LLMs) re-
main heuristic, lacking formal characterization of their fundamental computa-
tional limits. We establish the first complexity-theoretic foundation for LLM un-
learning, revealing intrinsic tradeoffs between efficiency, precision, and regulatory
compliance. Our framework formalizes (ϵ, δ)-machine unlearning via measure-
theoretic alignment of retrained and unlearned model distributions, then proves
transformer-specific hardness results: exact unlearning is coNP-hard, while ap-
proximate unlearning requires Ω(T 1−o(1)) time under the Exponential Time Hy-
pothesis (ETH). We construct an optimal Recursive Sketch-and-Freeze pro-
tocol achieving these bounds through differential privacy duality and Kronecker-
product sketching. Crucially, we identify phase transitions in R´enyi unlearning
cost at critical model scales (n ≈ d log k). These results provide (1) theoretical
benchmarks for evaluating unlearning algorithms, (2) complexity-aware guide-
lines for AI regulation, and (3) mathematically grounded verification tools for
GDPR/CPRA compliance.
|
false |
Imitation with Neural Density Models
Imitation Learning Reinforcement Learning Density Estimation Density Model Maximum Entropy RL Mujoco
We propose a new framework for Imitation Learning (IL) via density estimation of the expert's occupancy measure followed by Maximum Occupancy Entropy Reinforcement Learning (RL) using the density as a reward. Our approach maximizes a non-adversarial model-free RL objective that provably lower bounds reverse Kullback–Leibler divergence between occupancy measures of the expert and imitator. We present a practical IL algorithm, Neural Density Imitation (NDI), which obtains state-of-the-art demonstration efficiency on benchmark control tasks.
|
false |
Generative Discovery of Relational Medical Entity Pairs
Knowledge Discovery Generative Modeling Medical Entity Pair
Online healthcare services can provide the general public with ubiquitous access to medical knowledge and reduce the information access cost for both individuals and societies. To promote these benefits, it is desired to effectively expand the scale of high-quality yet novel relational medical entity pairs that embody rich medical knowledge in a structured form. To fulfill this goal, we introduce a generative model called Conditional Relationship Variational Autoencoder (CRVAE), which can discover meaningful and novel relational medical entity pairs without the requirement of additional external knowledge. Rather than discriminatively identifying the relationship between two given medical entities in a free-text corpus, we directly model and understand medical relationships from diversely expressed medical entity pairs. The proposed model introduces the generative modeling capacity of variational autoencoder to entity pairs, and has the ability to discover new relational medical entity pairs solely based on the existing entity pairs. Beside entity pairs, relationship-enhanced entity representations are obtained as another appealing benefit of the proposed method. Both quantitative and qualitative evaluations on real-world medical datasets demonstrate the effectiveness of the proposed method in generating relational medical entity pairs that are meaningful and novel.
|
false |
Not All Memories are Created Equal: Learning to Expire
expire long attention memory transformers
Attention mechanisms have shown promising results in sequence modeling tasks that require long-term memory. Recent work has investigated mechanisms to reduce the computational cost of preserving and storing the memories. However, not all content in the past is equally important to remember. We propose Expire-Span, a method that learns to retain the most important information and
expire the irrelevant information. This enables Transformers to scale to attend to tens of thousands of previous timesteps efficiently, as not all hidden states from previous timesteps are preserved. We demonstrate that Expire-Span can help models identify and retain critical information and show it can achieve state of the art results on long-context language modeling, reinforcement learning, and algorithmic tasks. Finally, we show that Expire-Span can scale to memories that are tens of thousands in size, which is helpful on incredibly long context tasks such as character-level PG-19 and a frame-by-frame moving objects task.
|
true |
GANs Can Play Lottery Tickets Too
lottery tickets GAN compression generative adversarial networks
Deep generative adversarial networks (GANs) have gained growing popularity in numerous scenarios, while usually suffer from high parameter complexities for resource-constrained real-world applications. However, the compression of GANs has less been explored. A few works show that heuristically applying compression techniques normally leads to unsatisfactory results, due to the notorious training instability of GANs. In parallel, the lottery ticket hypothesis shows prevailing success on discriminative models, in locating sparse matching subnetworks capable of training in isolation to full model performance. In this work, we for the first time study the existence of such trainable matching subnetworks in deep GANs. For a range of GANs, we certainly find matching subnetworks at $67\%$-$74\%$ sparsity. We observe that with or without pruning discriminator has a minor effect on the existence and quality of matching subnetworks, while the initialization weights used in the discriminator plays a significant role. We then show the powerful transferability of these subnetworks to unseen tasks. Furthermore, extensive experimental results demonstrate that our found subnetworks substantially outperform previous state-of-the-art GAN compression approaches in both image generation (e.g. SNGAN) and image-to-image translation GANs (e.g. CycleGAN). Codes available at https://github.com/VITA-Group/GAN-LTH.
|
true |
Witches' Brew: Industrial Scale Data Poisoning via Gradient Matching
Data Poisoning ImageNet Large-scale Gradient Alignment Security Backdoor Attacks from-scratch clean-label
Data Poisoning attacks modify training data to maliciously control a model trained on such data.
In this work, we focus on targeted poisoning attacks which cause a reclassification of an unmodified test image and as such breach model integrity. We consider a
particularly malicious poisoning attack that is both ``from scratch" and ``clean label", meaning we analyze an attack that successfully works against new, randomly initialized models, and is nearly imperceptible to humans, all while perturbing only a small fraction of the training data.
Previous poisoning attacks against deep neural networks in this setting have been limited in scope and success, working only in simplified settings or being prohibitively expensive for large datasets.
The central mechanism of the new attack is matching the gradient direction of malicious examples. We analyze why this works, supplement with practical considerations. and show its threat to real-world practitioners, finding that it is the first poisoning method to cause targeted misclassification in modern deep networks trained from scratch on a full-sized, poisoned ImageNet dataset.
Finally we demonstrate the limitations of existing defensive strategies against such an attack, concluding that data poisoning is a credible threat, even for large-scale deep learning systems.
|
true |
Learning to Set Waypoints for Audio-Visual Navigation
visual navigation audio visual learning embodied vision
In audio-visual navigation, an agent intelligently travels through a complex, unmapped 3D environment using both sights and sounds to find a sound source (e.g., a phone ringing in another room). Existing models learn to act at a fixed granularity of agent motion and rely on simple recurrent aggregations of the audio observations. We introduce a reinforcement learning approach to audio-visual navigation with two key novel elements: 1) waypoints that are dynamically set and learned end-to-end within the navigation policy, and 2) an acoustic memory that provides a structured, spatially grounded record of what the agent has heard as it moves. Both new ideas capitalize on the synergy of audio and visual data for revealing the geometry of an unmapped space. We demonstrate our approach on two challenging datasets of real-world 3D scenes, Replica and Matterport3D. Our model improves the state of the art by a substantial margin, and our experiments reveal that learning the links between sights, sounds, and space is essential for audio-visual navigation.
|
true |
Learning time-dependent PDE via graph neural networks and deep operator network for robust accuracy on irregular grids
Physical simulations Graph neural network Message passing neural PDE solvers Deep operator network DeepONet
There has been growing interest in models that learn the operator from the parameters of a partial differential equation (PDE) to the corresponding solutions. Deep Operator Network (DeepONet) and Fourier Neural operator, among other models, have been designed with structures suitable for handling functions as inputs and outputs, enabling real-time predictions as surrogate models for solution operators. There has also been significant progress in the research on surrogate models based on graph neural networks (GNNs), specifically targeting the dynamics in time-dependent PDEs. In this paper, we propose GraphDeepONet, an autoregressive model based on GNNs, to effectively adapt DeepONet, which is well-known for successful operator learning. GraphDeepONet exhibits robust accuracy in predicting solutions compared to existing GNN-based PDE solver models. It maintains consistent performance even on irregular grids, leveraging the advantages inherited from DeepONet and enabling predictions on arbitrary grids. Additionally, unlike traditional DeepONet and its variants, GraphDeepONet enables time extrapolation for time-dependent PDE solutions.
|
false |
On Evaluating Explainability Algorithms
interpretability Deep Learning
A plethora of methods attempting to explain predictions of black-box models have been proposed by the Explainable Artificial Intelligence (XAI) community. Yet, measuring the quality of the generated explanations is largely unexplored, making quantitative comparisons non-trivial. In this work, we propose a suite of multifaceted metrics that enables us to objectively compare explainers based on the correctness, consistency, as well as the confidence of the generated explanations. These metrics are computationally inexpensive, do not require model-retraining and can be used across different data modalities. We evaluate them on common explainers such as Grad-CAM, SmoothGrad, LIME and Integrated Gradients. Our experiments show that the proposed metrics reflect qualitative observations reported in earlier works.
|
false |
Improving Self-supervised Pre-training via a Fully-Explored Masked Language Model
representation learning natural language processing
Masked Language Model (MLM) framework has been widely adopted for self-supervised language pre-training. In this paper, we argue that randomly sampled masks in MLM would lead to undesirably large gradient variance. Thus, we theoretically quantify the gradient variance via correlating the gradient covariance with the Hamming distance between two different masks (given a certain text sequence). To reduce the variance due to the sampling of masks, we propose a fully-explored masking strategy, where a text sequence is divided into a certain number of non-overlapping segments. Thereafter, the tokens within one segment are masked for training. We prove, from a theoretical perspective, that the gradients derived from this new masking schema have a smaller variance and can lead to more efficient self-supervised training. We conduct extensive experiments on both continual pre-training and general pre-training from scratch. Empirical results confirm that this new masking strategy can consistently outperform standard random masking. Detailed efficiency analysis and ablation studies further validate the advantages of our fully-explored masking strategy under the MLM framework.
|
true |
Lipschitz Recurrent Neural Networks
recurrent neural networks dynamical systems differential equations
Viewing recurrent neural networks (RNNs) as continuous-time dynamical systems, we propose a recurrent unit that describes the hidden state's evolution with two parts: a well-understood linear component plus a Lipschitz nonlinearity. This particular functional form facilitates stability analysis of the long-term behavior of the recurrent unit using tools from nonlinear systems theory. In turn, this enables architectural design decisions before experimentation. Sufficient conditions for global stability of the recurrent unit are obtained, motivating a novel scheme for constructing hidden-to-hidden matrices. Our experiments demonstrate that the Lipschitz RNN can outperform existing recurrent units on a range of benchmark tasks, including computer vision, language modeling and speech prediction tasks. Finally, through Hessian-based analysis we demonstrate that our Lipschitz recurrent unit is more robust with respect to input and parameter perturbations as compared to other continuous-time RNNs.
|
false |
Generalized Universal Approximation for Certified Networks
adversarial deep learning neural network verification interval analysis
To certify safety and robustness of neural networks, researchers have successfully applied abstract interpretation, primarily using interval bound propagation. To understand the power of interval bounds, we present the abstract universal approximation (AUA) theorem, a generalization of the recent result by Baader et al. (2020) for ReLU networks to a large class of neural networks. The AUA theorem states that for any continuous function $f$, there exists a neural network that (1) approximates $f$ (universal approximation) and (2) whose interval bounds are an arbitrarily close approximation of the set semantics of $f$. The network may be constructed using any activation function from a rich class of functions---sigmoid, tanh, ReLU, ELU, etc.---making our result quite general. The key implication of the AUA theorem is that there always exists certifiably robust neural networks, which can be constructed using a wide range of activation functions.
|
true |
Fully Unsupervised Diversity Denoising with Convolutional Variational Autoencoders
Diversity denoising Unsupervised denoising Variational Autoencoders Noise model
Deep Learning based methods have emerged as the indisputable leaders for virtually all image restoration tasks. Especially in the domain of microscopy images, various content-aware image restoration (CARE) approaches are now used to improve the interpretability of acquired data. Naturally, there are limitations to what can be restored in corrupted images, and like for all inverse problems, many potential solutions exist, and one of them must be chosen. Here, we propose DivNoising, a denoising approach based on fully convolutional variational autoencoders (VAEs), overcoming the problem of having to choose a single solution by predicting a whole distribution of denoised images. First we introduce a principled way of formulating the unsupervised denoising problem within the VAE framework by explicitly incorporating imaging noise models into the decoder. Our approach is fully unsupervised, only requiring noisy images and a suitable description of the imaging noise distribution. We show that such a noise model can either be measured, bootstrapped from noisy data, or co-learned during training. If desired, consensus predictions can be inferred from a set of DivNoising predictions, leading to competitive results with other unsupervised methods and, on occasion, even with the supervised state-of-the-art. DivNoising samples from the posterior enable a plethora of useful applications. We are (i) showing denoising results for 13 datasets, (ii) discussing how optical character recognition (OCR) applications can benefit from diverse predictions, and are (iii) demonstrating how instance cell segmentation improves when using diverse DivNoising predictions.
|
false |
Zero-Shot Recognition through Image-Guided Semantic Classification
zero-shot learning visual-semantic embedding deep learning
We present a new visual-semantic embedding method for generalized zero-shot learning. Existing embedding-based methods aim to learn the correspondence between an image classifier (visual representation) and its class prototype (semantic representation) for each class. Inspired by the binary relevance method for multi-label classification, we learn the mapping between an image and its semantic classifier. Given an input image, the proposed Image-Guided Semantic Classification (IGSC) method creates a label classifier, being applied to all label embeddings to determine whether a label belongs to the input image. Therefore, a semantic classifier is image conditioned and is generated during inference. We also show that IGSC is a unifying framework for two state-of-the-art deep-embedding methods. We validate our approach with four standard benchmark datasets.
|
true |
De-biasing Weakly Supervised Learning by Regularizing Prediction Entropy
weakly learning prediction entropy entropy regularization prediction effect inexact class labels data distributions specific subclass training set
We explore the effect of regularizing prediction entropy in a weakly supervised setting with inexact class labels. When underlying data distributions are biased toward a specific subclass, we hypothesize that entropy regularization can be used to bootstrap a training set that mitigates this bias. We conduct experiments over multiple datasets under supervision of an oracle and in a semi-supervised setting finding substantial reductions in training set bias capable of decreasing test error rate. These findings suggest entropy regularization as a promising approach to de-biasing weakly supervised learning systems.
|
true |
The 37 Implementation Details of Proximal Policy Optimization
proximal-policy-optimization reproducibility reinforcement-learning implementation-details code-level-optimizations tutorial
Proximal policy optimization (PPO) has become one of the most popular deep reinforcement learning (DRL) algorithms. Yet, reproducing the PPO's results has been challenging in the community. While recent works conducted ablation studies to provide insight on PPO's implementation details, these works are not structured as tutorials and only focus on details concerning robotics tasks. As a result, reproducing PPO from scratch can become a daunting experience. Instead of introducing additional improvements, or doing further ablation studies, this blog post takes a step back and focuses on delivering a thorough reproduction of PPO in all accounts, as well as aggregating, documenting, and cataloging its most salient implementation details. This blog post also points out software engineering challenges in PPO and further efficiency improvement via the accelerated vectorized environments. With these, we believe this blog post will help people understand PPO faster and better, facilitating customization and research upon this versatile RL algorithm.
|
false |
Bounded Myopic Adversaries for Deep Reinforcement Learning Agents
deep reinforcement learning adversarial
Adversarial attacks against deep neural networks have been widely studied. Adversarial examples for deep reinforcement learning (DeepRL) have significant security implications, due to the deployment of these algorithms in many application domains. In this work we formalize an optimal myopic adversary for deep reinforcement learning agents. Our adversary attempts to find a bounded perturbation of the state which minimizes the value of the action taken by the agent. We show with experiments in various games in the Atari environment that our attack formulation achieves significantly larger impact as compared to the current state-of-the-art. Furthermore, this enables us to lower the bounds by several orders of magnitude on the perturbation needed to efficiently achieve significant impacts on DeepRL agents.
|
false |
Implementations that Matter in Cooperative Multi-Agent Reinforcement Learning
multi-agent reinforcement-learning experimental techniques monotonicity
Multi-agent reinforcement learning would like to implement some techniques that are known to have superior improvements in reinforcement learning. However, it is still unclear which of these techniques are complementary and can be fruitfully combined. This post dives into some proper experimental techniques that are controversial in multi-agent settings and gives out some advice to enhance the performance of MARL algorithms.
|
false |
Machine Learning Algorithms for Data Labeling: An Empirical Evaluation
Data Labeling Empirical Evaluation Active Machine Learning Semi-Supervised Learning
The lack of labeled data is a major problem in both research and industrial settings since obtaining labels is often an expensive and time-consuming activity. In the past years, several machine learning algorithms were developed to assist and perform automated labeling in partially labeled datasets. While many of these algorithms are available in open-source packages, there is no research that investigates how these algorithms compare to each other in different types of datasets and with different percentages of available labels. To address this problem, this paper empirically evaluates and compares seven algorithms for automated labeling in terms of accuracy. We investigate how these algorithms perform in six different and well-known datasets with three different types of data, images, texts, and numerical values. We evaluate these algorithms under two different experimental conditions, with 10\% and 50\% labels of available labels in the dataset. Each algorithm, in each dataset for each experimental condition, is evaluated independently ten times with different random seeds. The results are analyzed and the algorithms are compared utilizing a Bayesian Bradley-Terry model. The results indicate that while the algorithms label spreading with K-nearest neighbors perform better in the aggregated results, the active learning algorithms query by instance QBC and query instance uncertainty sample perform better when there is only 10\% of labels available. These results can help machine learning practitioners in choosing optimal machine learning algorithms to label their data.
|
false |
Joint Perception and Control as Inference with an Object-based Implementation
model-based reinforcement learning perception modeling object-based reinforcement learning
Existing model-based reinforcement learning methods often study perception modeling and decision making separately. We introduce joint Perception and Control as Inference (PCI), a general framework to combine perception and control for partially observable environments through Bayesian inference. Based on the fact that object-level inductive biases are critical in human perceptual learning and reasoning, we propose Object-based Perception Control (OPC), an instantiation of PCI which manages to facilitate control using automatic discovered object-based representations. We develop an unsupervised end-to-end solution and analyze the convergence of the perception model update. Experiments in a high-dimensional pixel environment demonstrate the learning effectiveness of our object-based perception control approach. Specifically, we show that OPC achieves good perceptual grouping quality and outperforms several strong baselines in accumulated rewards.
|
true |
Neural networks with late-phase weights
weights neural networks sgd learning weights neural networks successful variant stochastic gradient descent solutions subset
The largely successful method of training neural networks is to learn their weights using some variant of stochastic gradient descent (SGD). Here, we show that the solutions found by SGD can be further improved by ensembling a subset of the weights in late stages of learning. At the end of learning, we obtain back a single model by taking a spatial average in weight space. To avoid incurring increased computational costs, we investigate a family of low-dimensional late-phase weight models which interact multiplicatively with the remaining parameters. Our results show that augmenting standard models with late-phase weights improves generalization in established benchmarks such as CIFAR-10/100, ImageNet and enwik8. These findings are complemented with a theoretical analysis of a noisy quadratic problem which provides a simplified picture of the late phases of neural network learning.
|
true |
Orchestrating Tool Ecosystem of Drug Discovery with Intention-Aware LLM Agents
LLM agents drug discovery
Fragmented tools and models and complex decision-making with incomplete and heterogeneous information often hinder the drug discovery process. Large Language Models offer promising capabilities in commonsense reasoning and tool integration, yet their application in drug discovery remains constrained by challenges such as being incapable of handling large tool space, limited planning capabilities based on scientific intentions, and unscalable evaluation. We introduce GenieAgent, a drug discovery agent that integrates a wide range of molecule design models and bridges the user intentions to concrete actions by navigating the large skill ecosystem. By unifying disparate tools under a single natural language interface, GenieAgent enables cross-tool reasoning and supports complex scientific workflows. We also propose an evaluation framework simulating drug discovery conversations, based on real-world experiments. A large-scale assessment, validated by expert annotations, demonstrates that GenieAgent reliably meets the majority of molecular engineers' needs with high scientific accuracy and robustness.
|
false |
Concentric Spherical GNN for 3D Representation Learning
spherical cnn GNN graph convolution rotation equivariance 3D
Learning 3D representations that generalize well to arbitrarily oriented inputs is a challenge of practical importance in applications varying from computer vision to physics and chemistry.
We propose a novel multi-resolution convolutional architecture for learning over concentric spherical feature maps, of which the single sphere representation is a special case.
Our hierarchical architecture is based on alternatively learning to incorporate both intra-sphere and inter-sphere information.
We show the applicability of our method for two different types of 3D inputs, mesh objects, which can be regularly sampled, and point clouds, which are irregularly distributed.
We also propose an efficient mapping of point clouds to concentric spherical images using radial basis functions, thereby bridging spherical convolutions on grids with general point clouds.
We demonstrate the effectiveness of our approach in achieving state-of-the-art performance on 3D classification tasks with rotated data.
|
true |
Variational autoencoders trained with q-deformed lower bounds
variational autoencoders ELBO IWAE q-deformed logarithm tight bounds
Variational autoencoders (VAEs) have been successful at learning a low-dimensional manifold from high-dimensional data with complex dependencies. At their core, they consist of a powerful Bayesian probabilistic inference model, to capture the salient features of the data. In training, they exploit the power of variational inference, by optimizing a lower bound on the model evidence. The latent representation and the performance of VAEs are heavily influenced by the type of bound used as a cost function. Significant research work has been carried out into the development of tighter bounds than the original ELBO, to more accurately approximate the true log-likelihood. By leveraging the q-deformed logarithm in the traditional lower bounds, ELBO and IWAE, and the upper bound CUBO, we bring contributions to this direction of research. In this proof-of-concept study, we explore different ways of creating these q-deformed bounds that are tighter than the classical ones and we show improvements in the performance of such VAEs on the binarized MNIST dataset.
|
true |
Privately Learning from Graphs with Applications in Fine-tuning Large Pretrained Models
differential privacy relational learning private learning language models fine-tuning
Graphs offer unique insights into relationships and interactions between entities, complementing data modalities like text, images, and videos. By incorporating relational information from graph data, AI models can extend their capabilities beyond traditional tasks. However, relational data in sensitive domains such as finance and healthcare often contain private information, making privacy preservation crucial. Existing privacy-preserving methods, such as DP-SGD, which rely on gradient decoupling assumptions, are not well-suited for relational learning due to the inherent dependencies between coupled training samples. To address this challenge, we propose a privacy-preserving relational learning pipeline that decouples dependencies in sampled relations during training, ensuring differential privacy through a tailored application of DP-SGD. We apply this method to fine-tune large language models (LLMs) on sensitive graph data, and tackle the associated computational complexities. Our approach is evaluated on LLMs of varying sizes (e.g., BERT, Llama2) using real-world relational data from four text-attributed graphs. The results demonstrate significant improvements in relational learning tasks, all while maintaining robust privacy guarantees during training. Additionally, we explore the trade-offs between privacy, utility, and computational efficiency, offering insights into the practical deployment of our approach. Code is available at https://github.com/Graph-COM/PvGaLM.
|
true |
Annealed Importance Sampling meets Score Matching
importance sampling monte carlo annealing score matching
Annealed Importance Sampling (AIS) is one of the most effective methods for marginal likelihood estimation. It relies on a sequence of distributions interpolating between a tractable initial distribution and the posterior of interest which we simulate from approximately using a non-homogeneous Markov chain. To obtain an importance sampling (IS) estimate of the marginal likelihood, AIS introduces an extended target distribution to reweight the Markov chain proposal. While much effort has been devoted to improving the proposal distribution used by AIS by changing the intermediate distributions and corresponding Markov kernels, an underappreciated issue is that AIS uses an convenient but suboptimal extended target distribution which can hinder its performance. We leverage here recent progress in score-based generative modeling to learn the optimal extended target distribution for a given AIS proposal using score matching ideas. We demonstrate this novel differentiable AIS procedure on a number of synthetic benchmark distributions and a normalizing flow target.
|
false |
Nonconvex Continual Learning with Episodic Memory
continual learning nonconvex optimization
Continual learning aims to prevent catastrophic forgetting while learning a new task without accessing data of previously learned tasks.
The memory for such learning scenarios build a small subset of the data for previous tasks and is used in various ways such as quadratic programming and sample selection.
Current memory-based continual learning algorithms are formulated as a constrained optimization problem and rephrase constraints as a gradient-based approach.
However, previous works have not provided the theoretical proof on convergence to previously learned tasks.
In this paper, we propose a theoretical convergence analysis of continual learning based on stochastic gradient descent method.
Our method, nonconvex continual learning (NCCL), can achieve the same convergence rate when the proposed catastrophic forgetting term is suppressed at each iteration.
We also show that memory-based approaches have an inherent problem of overfitting to memory, which degrades the performance on previously learned tasks, namely catastrophic forgetting.
We empirically demonstrate that NCCL successfully performs continual learning with episodic memory by scaling learning rates adaptive to mini-batches on several image classification tasks.
|
true |
Influence Estimation for Generative Adversarial Networks
influence generative adversarial networks data cleansing
Identifying harmful instances, whose absence in a training dataset improves model performance, is important for building better machine learning models.
Although previous studies have succeeded in estimating harmful instances under supervised settings, they cannot be trivially extended to generative adversarial networks (GANs).
This is because previous approaches require that (i) the absence of a training instance directly affects the loss value and that (ii) the change in the loss directly measures the harmfulness of the instance for the performance of a model.
In GAN training, however, neither of the requirements is satisfied.
This is because, (i) the generator’s loss is not directly affected by the training instances as they are not part of the generator's training steps, and (ii) the values of GAN's losses normally do not capture the generative performance of a model.
To this end, (i) we propose an influence estimation method that uses the Jacobian of the gradient of the generator's loss with respect to the discriminator’s parameters (and vice versa) to trace how the absence of an instance in the discriminator’s training affects the generator’s parameters, and (ii) we propose a novel evaluation scheme, in which we assess harmfulness of each training instance on the basis of how GAN evaluation metric (e.g., inception score) is expected to change due to the removal of the instance.
We experimentally verified that our influence estimation method correctly inferred the changes in GAN evaluation metrics.
We also demonstrated that the removal of the identified harmful instances effectively improved the model’s generative performance with respect to various GAN evaluation metrics.
|
true |
Robust Reinforcement Learning for Autonomous Driving
Neural networks Deep reinforcement learning Actor-critic model Autonomous driving Carla simulator
Autonomous driving is still considered as an “unsolved problem” given its inherent important variability and that many processes associated with its development like vehicle control and scenes recognition remain open issues. Despite reinforcement learning algorithms have achieved notable results in games and some robotic manipulations, this technique has not been widely scaled up to the more challenging real world applications like autonomous driving. In this work, we propose a deep reinforcement learning (RL) algorithm embedding an actor critic architecture with multi-step returns to achieve a better robustness of the agent learning strategies when acting in complex and unstable environments. The experiment is conducted with Carla simulator offering a customizable and realistic urban driving conditions. The developed deep actor RL guided by a policy-evaluator critic distinctly surpasses the performance of a standard deep RL agent.
|
false |
Inductive Biases for Object-Centric Representations of Complex Textures
object-centric learning structured representations unsupervised learning deep learning
Understanding which inductive biases could be useful for unsupervised models to learn object-centric representations of natural scenes is challenging. Here, we use neural style transfer to generate datasets where objects have complex textures. Our main finding is that a model that reconstructs both the shape and visual appearance of each object from its representation achieves correct separation of the objects and learns useful object representations more reliably.
|
false |
Learning to Noise: Application-Agnostic Data Sharing with Local Differential Privacy
Differential Privacy Representation Learning Variational Inference Generative Modelling
In recent years, the collection and sharing of individuals’ private data has become commonplace in many industries. Local differential privacy (LDP) is a rigorous approach which uses a randomized algorithm to preserve privacy even from the database administrator, unlike the more standard central differential privacy. For LDP, when applying noise directly to high-dimensional data, the level of noise required all but entirely destroys data utility. In this paper we introduce a novel, application-agnostic privatization mechanism that leverages representation learning to overcome the prohibitive noise requirements of direct methods, while maintaining the strict guarantees of LDP. We further demonstrate that data privatized with this mechanism can be used to train machine learning algorithms. Applications of this model include private data collection, private novel-class classification, and the augmentation of clean datasets with additional privatized features. We achieve significant gains in performance on downstream classification tasks relative to benchmarks that noise the data directly, which are state-of-the-art in the context of application-agnostic LDP mechanisms for high-dimensional data sharing tasks.
|
false |
On Nondeterminism and Instability in Neural Network Optimization
Nondeterminism Instability
Optimization nondeterminism causes uncertainty when improving neural networks, with small changes in performance difficult to discern from run-to-run variability. While uncertainty can be reduced by training multiple copies of a model with different random seeds, doing so is time-consuming, costly, and makes reproducibility challenging. Despite this, little attention has been paid towards establishing an understanding of this problem. In this work, we establish an experimental protocol for understanding the effect of optimization nondeterminism on model diversity, which allows us to study the independent effects of a variety of sources of nondeterminism. Surprisingly, we find that each source of nondeterminism all have similar effects on multiple measures of model diversity. To explain this intriguing fact, we examine and identify the instability of model training, when taken as an end-to-end procedure, as the key determinant. We show that even one-bit changes in initial model parameters result in models that converge to vastly different values. Last, we demonstrate that recent methods in accelerated model ensembling hold promise for reducing the effects of instability on run-to-run variability.
|
true |
Adversarial Mixup Resynthesizers
autoencoders adversarial GANs
In this paper, we explore new approaches to combining information encoded within the learned representations of autoencoders. We explore models that are capable of combining the attributes of multiple inputs such that a resynthesised output is trained to fool an adversarial discriminator for real versus synthesised data. Furthermore, we explore the use of such an architecture in the context of semi-supervised learning, where we learn a mixing function whose objective is to produce interpolations of hidden states, or masked combinations of latent representations that are consistent with a conditioned class label. We show quantitative and qualitative evidence that such a formulation is an interesting avenue of research.
|
true |
Disentanglement and Generalization Under Correlation Shifts
disentanglement correlation mutual information conditional mutual information subspace independence
Correlations between factors of variation are prevalent in real-world data. However, often such correlations are not robust (e.g., they may change between domains, datasets, or applications) and we wish to avoid exploiting them. Disentanglement methods aim to learn representations which capture different factors of variation in latent subspaces. A common approach involves minimizing the mutual information between latent subspaces, such that each encodes a single underlying attribute. However, this fails when attributes are correlated. We solve this problem by enforcing independence between subspaces conditioned on the available attributes, which allows us to remove only dependencies that are not due to the correlation structure present in the training data. We achieve this via an adversarial approach to minimize the conditional mutual information (CMI) between subspaces with respect to categorical variables. We first show theoretically that CMI minimization is a good objective for robust disentanglement on linear problems with Gaussian data. We then apply our method on real-world datasets based on MNIST and CelebA, and show that it yields models that are disentangled and robust under correlation shift, including in weakly supervised settings.
|
true |
Scalable Fingerprinting of Large Language Models
Fingerprinting
Model fingerprinting has emerged as a powerful tool for model owners to identify their shared model given API access. However, to lower false discovery rate, fight fingerprint leakage, and defend against coalitions of model users attempting to bypass detection, we argue that scaling up the number of fingerprints one can embed into a model is critical. Hence, we pose Scalability as a crucial requirement for good fingerprinting schemes.
We experiment with fingerprint design at larger scales than previously considered,
and propose a new method, dubbed Perinucleus sampling, to generate scalable, persistent, and harmless fingerprints. We demonstrate that this scheme can add 24,576 fingerprints to a Llama-3.1-8B model --- two orders of magnitude more than existing schemes --- without degrading the model's utility. Our inserted fingerprints persist even after supervised fine-tuning on other data. We further describe security risks for fingerprinting, and theoretically and empirically show how a scalable fingerprinting scheme like ours can help mitigate these risks.
|
false |
Align-RUDDER: Learning From Few Demonstrations by Reward Redistribution
demonstrations rudder complex tasks high rewards learning reward redistribution reward delay rewards lstm model
Reinforcement Learning algorithms require a large number of samples to solve complex tasks with sparse and delayed rewards.
Complex tasks are often hierarchically composed of sub-tasks.
A step in the Q-function indicates solving a sub-task, where the expectation of the return increases.
RUDDER identifies these steps and then redistributes reward to them, thus immediately giving reward if sub-tasks are solved.
Since the delay of rewards is reduced, learning is considerably sped up.
However, for complex tasks, current exploration strategies struggle with discovering episodes with high rewards.
Therefore, we assume that episodes with high rewards are given as demonstrations and do not have to be discovered by exploration.
Typically the number of demonstrations is small and RUDDER's LSTM model does not learn well.
Hence, we introduce Align-RUDDER, which is RUDDER with two major modifications.
First, Align-RUDDER assumes that episodes with high rewards are given as demonstrations,
replacing RUDDER’s safe exploration and lessons replay buffer.
Second, we substitute RUDDER’s LSTM model by a profile model that is obtained from multiple sequence alignment of demonstrations.
Profile models can be constructed from as few as two demonstrations.
Align-RUDDER inherits the concept of reward redistribution, which speeds up learning by reducing the delay of rewards.
Align-RUDDER outperforms competitors on complex artificial tasks with delayed reward and few demonstrations.
On the MineCraft ObtainDiamond task, Align-RUDDER is able to mine a diamond, though not frequently.
|
false |
Towards Unifying Interpretability and Control: Evaluation via Intervention
mechanistic interpretability evaluation explainability safety
With the growing complexity and capability of large language models, a need to understand model reasoning has emerged, often motivated by an underlying goal of controlling and aligning models.
While numerous interpretability and steering methods have been proposed as solutions, they are typically designed either for understanding or for control, seldom addressing both. Additionally, the lack of standardized applications, motivations, and evaluation metrics makes it difficult to assess methods' practical utility and efficacy.
To address the aforementioned issues, we argue that intervention is a fundamental goal of interpretability and introduce success criteria to evaluate how well methods can control model behavior through interventions. To evaluate existing methods for this ability, we unify and extend four popular interpretability methods—sparse autoencoders, logit lens, tuned lens, and probing—into an abstract encoder-decoder framework, enabling interventions on interpretable features that can be mapped back to latent representations to control model outputs.
We introduce two new evaluation metrics: intervention success rate and coherence-intervention tradeoff, designed to measure the accuracy of explanations and their utility in controlling model behavior. Our findings reveal that (1) while current methods allow for intervention, their effectiveness is inconsistent across features and models, (2) lens-based methods outperform SAEs and probes in achieving simple, concrete interventions, and (3) mechanistic interventions often compromise model coherence, underperforming simpler alternatives, such as prompting, and highlighting a critical shortcoming of current interpretability approaches in applications requiring control.
|
false |
Test-Time Adaptation and Adversarial Robustness
test-time adaptation adversarial robustness unsupervised domain adaptation threat model maximin game
This paper studies test-time adaptation in the context of adversarial robustness. We formulate an adversarial threat model for test-time adaptation, where the defender may have a unique advantage as the adversarial game becomes a maximin game, instead of a minimax game as in the classic adversarial robustness threat model. We then study whether the maximin threat model admits more ``good solutions'' than the minimax threat model, and is thus \emph{strictly weaker}. For this purpose, we first present a provable separation between the two threat models in a natural Gaussian data model. For deep learning, while we do not have a proof, we propose a candidate, Domain Adversarial Neural Networks (${\sf DANN}$), an algorithm designed for unsupervised domain adaptation, by showing that it provides nontrivial robustness in the test-time maximin threat model against strong transfer attacks and adaptive attacks. This is somewhat surprising since ${\sf DANN}$ is not designed specifically for adversarial robustness (e.g., against norm-based attacks), and provides no robustness in the minimax model. Complementing these results, we show that recent data-oblivious test-time adaptations can be easily attacked even with simple transfer attacks. We conclude the paper with various future directions of studying adversarially robust test-time adaptation.
|
false |
Wiring Up Vision: Minimizing Supervised Synaptic Updates Needed to Produce a Primate Ventral Stream
computational neuroscience primate ventral stream convolutional neural networks biologically plausible learning
After training on large datasets, certain deep neural networks are surprisingly good models of the neural mechanisms of adult primate visual object recognition. Nevertheless, these models are poor models of the development of the visual system because they posit millions of sequential, precisely coordinated synaptic updates, each based on a labeled image. While ongoing research is pursuing the use of unsupervised proxies for labels, we here explore a complementary strategy of reducing the required number of supervised synaptic updates to produce an adult-like ventral visual stream (as judged by the match to V1, V2, V4, IT, and behavior). Such models might require less precise machinery and energy expenditure to coordinate these updates and would thus move us closer to viable neuroscientific hypotheses about how the visual system wires itself up. Relative to the current leading model of the adult ventral stream, we here demonstrate that the total number of supervised weight updates can be substantially reduced using three complementary strategies: First, we find that only 2% of supervised updates (epochs and images) are needed to achieve ~80% of the match to adult ventral stream. Second, by improving the random distribution of synaptic connectivity, we find that 54% of the brain match can already be achieved “at birth" (i.e. no training at all). Third, we find that, by training only ~5% of model synapses, we can still achieve nearly 80% of the match to the ventral stream. When these three strategies are applied in combination, we find that these new models achieve ~80% of a fully trained model's match to the brain, while using two orders of magnitude fewer supervised synaptic updates. These results reflect first steps in modeling not just primate adult visual processing during inference, but also how the ventral visual stream might be "wired up" by evolution (a model's "birth" state) and by developmental learning (a model's updates based on visual experience).
|
false |
Formal Language Constrained Markov Decision Processes
safe reinforcement learning formal languages constrained Markov decision process safety gym safety
In order to satisfy safety conditions, an agent may be constrained from acting freely. A safe controller can be designed a priori if an environment is well understood, but not when learning is employed. In particular, reinforcement learned (RL) controllers require exploration, which can be hazardous in safety critical situations. We study the benefits of giving structure to the constraints of a constrained Markov decision process by specifying them in formal languages as a step towards using safety methods from software engineering and controller synthesis. We instantiate these constraints as finite automata to efficiently recognise constraint violations. Constraint states are then used to augment the underlying MDP state and to learn a dense cost function, easing the problem of quickly learning joint MDP/constraint dynamics. We empirically evaluate the effect of these methods on training a variety of RL algorithms over several constraints specified in Safety Gym, MuJoCo, and Atari environments.
|
true |
Improving VAEs' Robustness to Adversarial Attack
deep generative models variational autoencoders robustness adversarial attack
Variational autoencoders (VAEs) have recently been shown to be vulnerable to adversarial attacks, wherein they are fooled into reconstructing a chosen target image. However, how to defend against such attacks remains an open problem. We make significant advances in addressing this issue by introducing methods for producing adversarially robust VAEs. Namely, we first demonstrate that methods proposed to obtain disentangled latent representations produce VAEs that are more robust to these attacks. However, this robustness comes at the cost of reducing the quality of the reconstructions. We ameliorate this by applying disentangling methods to hierarchical VAEs. The resulting models produce high--fidelity autoencoders that are also adversarially robust. We confirm their capabilities on several different datasets and with current state-of-the-art VAE adversarial attacks, and also show that they increase the robustness of downstream tasks to attack.
|
false |
Motif-Driven Contrastive Learning of Graph Representations
graph neural network self-supervised learning contrastive learning graph motif learning
Graph motifs are significant subgraph patterns occurring frequently in graphs, and they play important roles in representing the whole graph characteristics. For example, in the chemical domain, functional groups are motifs that can determine molecule properties. Mining and utilizing motifs, however, is a non-trivial task for large graph datasets. Traditional motif discovery approaches mostly rely on exact counting or statistical estimation, which are hard to scale for a large number of graphs with continuous and high-dimension features. In light of the significance and challenges of motif mining, we propose : MICRO-Graph: a framework for \underline{M}ot\underline{I}f-driven \underline{C}ontrastive lea\underline{R}ning \underline{O}f \underline{G}raph representations to: 1) pre-train Graph Neural Networks (GNNs) in a self-supervised manner to automatically extract graph motifs from large graph datasets; 2) leverage learned motifs to guide the contrastive learning of graph representations, which further benefit various graph downstream tasks. Specifically, given a graph dataset, a motif learner cluster similar and significant subgraphs into corresponding motif slots. Based on the learned motifs, a motif-guided subgraph segmenter is trained to generate more informative subgraphs, which are used to conduct graph-to-subgraph contrastive learning of GNNs. Our discovering strategy is to simutaneously do clustering and contrastive learning on dynamically sampled subgraphs. The clustering part pull together similar subgraphs across different whole graphs, as the contrastive part push away dissimilar ones. Meanwhile, our learnable sampler will generate subgraph samples better aligned with the discoverying procedure. By pre-training on ogbn-molhiv molecule dataset with our proposed MICRO-Graph, the pre-trained GNN model can enhance various chemical property prediction downstream tasks with scarce label by 2.0%, and significantly higher than other state-of-the-art self-supervised learning baselines.
|
false |
Rethinking Compressed Convolution Neural Network from a Statistical Perspective
Compressed Convolutional Neural Network Tensor Decomposition Sample Complexity Analysis
Many designs have recently been proposed to improve the model efficiency of convolutional neural networks (CNNs) at a fixed resource budget, while there is a lack of theoretical analysis to justify them. This paper first formulates CNNs with high-order inputs into statistical models, which have a special "Tucker-like" formulation. This makes it possible to further conduct the sample complexity analysis to CNNs as well as compressed CNNs via tensor decomposition. Tucker and CP decompositions are commonly adopted to compress CNNs in the literature. The low rank assumption is usually imposed on the output channels, which according to our study, may not be beneficial to obtain a computationally efficient model while a similar accuracy can be maintained. Our finding is further supported by ablation studies on CIFAR10, SVNH and UCF101 datasets.
|
false |
Motion Forecasting with Unlikelihood Training
contextual information unlikelihood motion forecasting model models motion essential safe intelligent decisions robotic applications
Motion forecasting is essential for making safe and intelligent decisions in robotic applications such as autonomous driving. State-of-the-art methods formulate it as a sequence-to-sequence prediction problem, which is solved in an encoder-decoder framework with a maximum likelihood estimation objective. In this paper, we show that the likelihood objective itself results in a model assigning too much probability to trajectories that are unlikely given the contextual information such as maps and states of surrounding agents. This is despite the fact that many state-of-the-art models do take contextual information as part of their input. We propose a new objective, unlikelihood training, which forces generated trajectories that conflicts with contextual information to be assigned a lower probability by our model. We demonstrate that our method can significantly improve state-of-art models’ performance on challenging real-world trajectory forecasting datasets (nuScenes and Argoverse) by 8% and reduce the standard deviation by up to 50%. The code will be made available.
|
true |
Regularized Inverse Reinforcement Learning
inverse reinforcement learning reward learning regularized markov decision processes reinforcement learning
Inverse Reinforcement Learning (IRL) aims to facilitate a learner’s ability to imitate expert behavior by acquiring reward functions that explain the expert’s decisions. Regularized IRLapplies strongly convex regularizers to the learner’s policy in order to avoid the expert’s behavior being rationalized by arbitrary constant rewards, also known as degenerate solutions. We propose tractable solutions, and practical methods to obtain them, for regularized IRL. Current methods are restricted to the maximum-entropy IRL framework, limiting them to Shannon-entropy regularizers, as well as proposing solutions that are intractable in practice. We present theoretical backing for our proposed IRL method’s applicability to both discrete and continuous controls, empirically validating our performance on a variety of tasks.
|
false |
Register Allocation using Graph Neural Networks
Register allocation graph coloring graph neural networks
The execution lifecycle of a program goes through various optimizations and code generation. Register allocation is critical part of optimization phase of compiler. Graph coloring is one of the NP hard problem and it applies to various applications such as register allocation. Deep Learning methods has various state-of-art techniques in domains like computer vision, natural language processing.
In this paper we propose a solution for register allocation problem of compiler using Deep Learning networks based on graphs called as graph neural networks (GNN). We start with basic idea that two connected nodes with temporaries will have different registers allocated.
|
false |
Connecting Sphere Manifolds Hierarchically for Regularization
Hierarchy Manifold Classification
This paper considers classification problems with hierarchically organized classes. We force the classifier (hyperplane) of each class to belong to a sphere manifold, whose center is the classifier of its super-class. Then, individual sphere manifolds are connected based on their hierarchical relations. Our technique replaces the last layer of a neural network by combining a spherical fully-connected layer with a hierarchical layer. This regularization is shown to improve the performance of widely used deep neural network architectures (ResNet and DenseNet) on publicly available datasets (CIFAR100, CUB200, Stanford dogs, Stanford cars, and Tiny-ImageNet).
|
false |
Detecting Misclassification Errors in Neural Networks with a Gaussian Process Model
Neural Network Classifier Error Detection AI safety
As neural network classifiers are deployed in real-world applications, it is crucial that their predictions are not just accurate, but trustworthy as well. One practical solution is to assign confidence scores to each prediction, then filter out low-confidence predictions. However, existing confidence metrics are not yet sufficiently reliable for this role. This paper presents a new framework that produces more reliable confidence scores for detecting misclassification errors. This framework, RED, calibrates the classifier's inherent confidence indicators and estimates uncertainty of the calibrated confidence scores using Gaussian Processes. Empirical comparisons with other confidence estimation methods on 125 UCI datasets demonstrate that this approach is effective. An experiment on a vision task with a large deep learning architecture further confirms that the method can scale up, and a case study involving out-of-distribution and adversarial samples shows potential of the proposed method to improve robustness of neural network classifiers more broadly in the future.
|
false |
Compositional Models: Multi-Task Learning and Knowledge Transfer with Modular Networks
modular networks transfer learning domain adaptation self-organization
Conditional computation and modular networks have been recently proposed for multitask learning and other problems as a way to decompose problem solving into multiple reusable computational blocks. We propose a novel fully-differentiable approach for learning modular networks.In our method, the modules can be invoked repeatedly and allow knowledge transfer to novel tasks by adjusting the order of computation. This allows soft weight sharing between tasks with only a small increase in the number of parameters. We show that our method leads to interpretable self-organization of modules in case of multi-task learning, transfer learning and domain adaptation while achieving competitive results on those tasks. From practical perspective, our approach allows to: (a) reuse existing modules for learning new task by adjusting the computation order, (b) use it for unsupervised multi-source domain adaptation to illustrate that adaptation to unseen data can be achieved by only manipulating the order of pretrained modules, (c) show how our approach can be used to increase accuracy of existing architectures for image classification tasks such as ImageNet, without any parameter increase, by reusing the same block multiple times.
|
false |
DEMI: Discriminative Estimator of Mutual Information
Mutual information estimation discriminative classification
Estimating mutual information between continuous random variables is often intractable and extremely challenging for high-dimensional data. Recent progress has leveraged neural networks to optimize variational lower bounds on mutual information. Although showing promise for this difficult problem, the variational methods have been theoretically and empirically proven to have serious statistical limitations: 1) many methods struggle to produce accurate estimates when the underlying mutual information is either low or high; 2) the resulting estimators may suffer from high variance. Our approach is based on training a classifier that provides the probability that a data sample pair is drawn from the joint distribution rather than from the product of its marginal distributions. Moreover, we establish a direct connection between mutual information and the average log odds estimate produced by the classifier on a test set, leading to a simple and accurate estimator of mutual information. We show theoretically that our method and other variational approaches are equivalent when they achieve their optimum, while our method sidesteps the variational bound. Empirical results demonstrate high accuracy of our approach and the advantages of our estimator in the context of representation learning.
|
false |
Evaluating Gender Bias in Natural Language Inference
Natural Language Inference Natural Language Understanding Natural Language Processing Gender Bias Societal Bias Bias Ethics Debiasing Techniques Data Augmentation
Gender-bias stereotypes have recently raised significant ethical concerns in natural language processing. However, progress in the detection and evaluation of gender-bias in natural language understanding through inference is limited and requires further investigation. In this work, we propose an evaluation methodology to measure these biases by constructing a probe task that involves pairing a gender-neutral premise against a gender-specific hypothesis. We use our probe task to investigate state-of-the-art NLI models on the presence of gender stereotypes using occupations. Our findings suggest that three models (BERT, RoBERTa, and BART) trained on MNLI and SNLI data-sets are significantly prone to gender-induced prediction errors. We also find that debiasing techniques such as augmenting the training dataset to ensure that it is a gender-balanced dataset can help reduce such bias in certain cases.
|
false |
SOAR: Second-Order Adversarial Regularization
Adversarial Robustness
Adversarial training is a common approach to improving the robustness of deep neural networks against adversarial examples. In this work, we propose a novel regularization approach as an alternative. To derive the regularizer, we formulate the adversarial robustness problem under the robust optimization framework and approximate the loss function using a second-order Taylor series expansion. Our proposed second-order adversarial regularizer (SOAR) is an upper bound based on the Taylor approximation of the inner-max in the robust optimization objective. We empirically show that the proposed method improves the robustness of networks against the $\ell_\infty$ and $\ell_2$ bounded perturbations on CIFAR-10 and SVHN.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.