39 citations · 158 across the 36 of their papers we have counts for
15 papers · 1 filter
Scaling Laws for Associative Memories
Vivien Cabannes, Elvis Dohmatob, Alberto Bietti
Learning arguably involves the discovery and memorization of abstract rules. The aim of this paper is to study associative memory mechanisms. Our model is based on high-dimensional…
Robust Linear Regression: Phase-Transitions and Precise Tradeoffs for General Norms
Elvis Dohmatob, Meyer Scetbon
In this paper, we investigate the impact of test-time adversarial attacks on linear regression models and determine the optimal level of robustness that any model can reach while m…
Robust Linear Regression: Gradient-descent, Early-stopping, and Beyond
Meyer Scetbon, Elvis Dohmatob
In this work we study the robustness to adversarial attacks, of early-stopping strategies on gradient-descent (GD) methods for linear regression. More precisely, we show that early…
Origins of Low-dimensional Adversarial Perturbations
Elvis Dohmatob, Chuan Guo, Morgane Goibert
In this paper, we initiate a rigorous study of the phenomenon of low-dimensional adversarial perturbations (LDAPs) in classification. Unlike the classical setting, these perturbati…
On the (Non-)Robustness of Two-Layer Neural Networks in Different Learning Regimes
Elvis Dohmatob, Alberto Bietti
Neural networks are known to be highly sensitive to adversarial examples. These may arise due to different factors, such as random initialization, or spurious correlations in the l…
Fundamental tradeoffs between memorization and robustness in random features and neural tangent regimes
Elvis Dohmatob
This work studies the (non)robustness of two-layer neural networks in various high-dimensional linearized regimes. We establish fundamental trade-offs between memorization and robu…
Implicit bias of any algorithm: bounding bias via margin
Elvis Dohmatob
Consider points in finite-dimensional euclidean space, each having one of two colors. Suppose there exists a separating hyperplane (identified with its unit no…
Classifier-independent Lower-Bounds for Adversarial Robustness
Elvis Dohmatob
We theoretically analyse the limits of robustness to test-time adversarial and noisy examples in classification. Our work focuses on deriving bounds which uniformly apply to all cl…
Learning disconnected manifolds: a no GANs land
Ugo Tanielian, Thibaut Issenhuth, Elvis Dohmatob +1
Typical architectures of Generative AdversarialNetworks make use of a unimodal latent distribution transformed by a continuous generator. Consequently, the modeled distribution alw…
On the Convergence of Approximate and Regularized Policy Iteration Schemes
Elena Smirnova, Elvis Dohmatob
Entropy regularized algorithms such as Soft Q-learning and Soft Actor-Critic, recently showed state-of-the-art performance on a number of challenging reinforcement learning (RL) ta…
Distributionally Robust Counterfactual Risk Minimization
Louis Faury, Ugo Tanielian, Flavian Vasile +2
This manuscript introduces the idea of using Distributionally Robust Optimization (DRO) for the Counterfactual Risk Minimization (CRM) problem. Tapping into a rich existing literat…
Distributionally Robust Reinforcement Learning
Elena Smirnova, Elvis Dohmatob, Jérémie Mary
Real-world applications require RL algorithms to act safely. During learning process, it is likely that the agent executes sub-optimal actions that may lead to unsafe/poor states o…
Deep Determinantal Point Processes
Mike Gartrell, Elvis Dohmatob, Jon Alberdi
Determinantal point processes (DPPs) have attracted significant attention as an elegant model that is able to capture the balance between quality and diversity within sets. DPPs ar…
Generalized No Free Lunch Theorem for Adversarial Robustness
Elvis Dohmatob
This manuscript presents some new impossibility results on adversarial robustness in machine learning, a very important yet largely open problem. We show that if conditioned on a c…
Continuation of Nesterov's Smoothing for Regression with Structured Sparsity in High-Dimensional Neuroimaging
Fouad Hadj-Selem, Tommy Lofstedt, Elvis Dohmatob +4
Predictive models can be used on high-dimensional brain images for diagnosis of a clinical condition. Spatial regularization through structured sparsity offers new perspectives in…