Publications by Paul Thierry Yves Rolland

Predicting in Uncertain Environments: Methods for Robust Machine Learning

One of the main goal of Artificial Intelligence is to develop models capable of providing valuable predictions in real-world environments. In particular, Machine Learning (ML) seeks to design such models by learning from examples coming from this same envi ...

EPFL2022

Score Matching Enables Causal Discovery of Nonlinear Additive Noise Models

Volkan Cevher, Paul Thierry Yves Rolland

This paper demonstrates how to recover causal graphs from the score of the data distribution in non-linear additive (Gaussian) noise models. Using score matching algorithms as a building block, we show how to design a new generation of scalable causal disc ...

JMLR-JOURNAL MACHINE LEARNING RESEARCH2022

Identifiability and Generalizability from Multiple Experts in Inverse Reinforcement Learning

Volkan Cevher, Paul Thierry Yves Rolland, Luca Viano, Boris Nikolov

While Reinforcement Learning (RL) aims to train an agent from a reward function in a given environment, Inverse Reinforcement Learning (IRL) seeks to recover the reward function from observing an expert’s behavior. It is well known that, in general, variou ...

2022

The Effect of the Intrinsic Dimension on the Generalization of Quadratic Classifiers

Paul Thierry Yves Rolland, Leello Tadesse Dadi, Fabian Ricardo Latorre Gomez

It has been recently observed that neural networks, unlike kernel methods, enjoy a reduced sample complexity when the distribution is isotropic (i.e., when the covariance matrix is the identity). We find that this sensitivity to the data distribution is no ...

2021

Double-Loop Unadjusted Langevin Algorithm

Volkan Cevher, Paul Thierry Yves Rolland, Ali Kavis, Armin Eftekhari

A well-known first-order method for sampling from log-concave probability distributions is the Unadjusted Langevin Algorithm (ULA). This work proposes a new annealing step-size schedule for ULA, which allows to prove new convergence guarantees for sampling ...

2020

Efficient Proximal Mapping of the 1-path-norm of Shallow Networks

Volkan Cevher, Paul Thierry Yves Rolland, Fabian Ricardo Latorre Gomez, Shaul Nadav Hallak

We demonstrate two new important properties of the 1-path-norm of shallow neural networks. First, despite its non-smoothness and non-convexity it allows a closed form proximal operator which can be efficiently computed, allowing the use of stochastic proxi ...

JMLR-JOURNAL MACHINE LEARNING RESEARCH2020

Lipschitz constant estimation for Neural Networks via sparse polynomial optimization

Volkan Cevher, Paul Thierry Yves Rolland, Fabian Ricardo Latorre Gomez

We introduce LiPopt, a polynomial optimization framework for computing increasingly tighter upper bounds on the Lipschitz constant of neural networks. The underlying optimization problems boil down to either linear (LP) or semidefinite (SDP) programming. W ...

2020

Robust Reinforcement Learning via Adversarial training with Langevin Dynamics

Volkan Cevher, Paul Thierry Yves Rolland, Ya-Ping Hsieh, Kamalaruban Parameswaran, Yu-Ting Huang, Cheng Shi

We introduce a sampling perspective to tackle the challenging task of training robust Reinforcement Learning (RL) agents. Leveraging the powerful Stochastic Gradient Langevin Dynamics, we present a novel, scalable two-player RL algorithm, which is a sampli ...

34th Conference on Neural Information Processing Systems (NeurIPS 2020)2020

Efficient learning of smooth probability functions from Bernoulli tests with guarantees

Volkan Cevher, Paul Thierry Yves Rolland, Ali Kavis

We study the fundamental problem of learning an unknown, smooth probability function via pointwise Bernoulli tests. We provide a scalable algorithm for efficiently solving this problem with rigorous guarantees. In particular, we prove the convergence rate ...

2019

High Dimensional Bayesian Optimization via Additive Models with Overlapping Groups

Volkan Cevher, Paul Thierry Yves Rolland, Jonathan Mark Scarlett, Ilija Bogunovic

Bayesian optimization (BO) is a popular technique for sequential black-box function optimization, with applications including parameter tuning, robotics, environmental monitoring, and more. One of the most important challenges in BO is the development of a ...

2018