Luca Viano (2026) ‒ LIONS ‐ EPFL

Research Interests:

Reinforcement Learning
Imitation Learning
Optimization

Biography

On July 17th, 2026 I successfully defended my PhD thesis. The thesis, entitled “Reward is not enough: Advances in reinforcement learning from demonstrations and preferences” was supervised by Professor Volkan Cevher.

I did my PhD studies at LIONS. Since September 2021, I was an ELLIS PhD Fellow supervised by Volkan Cevher and co-supervised by Gergely Neu. My first PhD year was kindly supported by the EPFL EDIC Excellence Fellowship.

Before that I worked as Data Scientist at Datapred, I obtained a MSc in Computational Science and Engineering from EPFL and a Bachelor in Engineering Physics from Politecnico di Torino, Italy. Cycling is my main hobby.

Multi-agent imitation learning with function approximation: Linear Markov games and beyond

L. Viano; T. Freihaut; E. Nevali; V. Cevher; M. Geist et al.

2026. 43rd International Conference on Machine Learning (ICML 2026), Seoul, South Korea, 2026-07-06 – 2026-07-11.

Research Interests:

Biography

Multi-agent imitation learning with function approximation: Linear Markov games and beyond

Aligning Large Language Models With Human Feedback: Mathematical foundations and algorithm design [Special Issue on the Mathematics of Deep Learning]

Convergence to Equilibrium of No-Regret Dynamics in Congestion Games

Reward is not enough: Advances in reinforcement learning fromdemonstrations and preferences

Multi-Step Alignment as Markov Games: An Optimistic Online Mirror Descent Approach with Convergence Guarantees

Learning Equilibria from Data: Provably Efficient Multi-Agent Imitation Learning

IL-SOAR : Imitation Learning with Soft Optimistic Actor cRitic

Best of Both Worlds: Regret Minimization versus Minimax Play

Adaptive Bilevel Optimization

Method for imaging a sample with an imaging device

Imitation Learning in Discounted Linear MDPs without exploration assumptions

Semi Bandit Dynamics in Congestion Games: Convergence to Nash Equilibrium and No-Regret Guarantees.

Alternation makes the adversary weaker in two-player games

What can online reinforcement learning with function approximation benefitfrom general coverage conditions

Identifiability and Generalizability from Multiple Experts in Inverse Reinforcement Learning