Alex Slivkins

2 accepted papers

2026

Bandit Social Leaning Dynamics with Exploration Episodes

Kiarash Banihashem, Collina, Alex Slivkins

ICML 2026poster

We study a stylized social learning dynamics where self-interested agents collectively follow a simple multi-armed bandit protocol. Each agent controls an "episode": a short sequence of consecutive decisions. Motivating applications include users repeatedly interacting with an AI, or repeatedly shop…

Cited by 0 Source ↗Scholar ↗

2024

Oracle-Efficient Pessimism: Offline Policy Optimization In Contextual Bandits

Lequn Wang, Akshay Krishnamurthy, Alex Slivkins

AISTATS 2024poster

We consider offline policy optimization (OPO) in contextual bandits, where one is given a fixed dataset of logged interactions. While pessimistic regularizers are typically used to mitigate distribution shift, prior implementations thereof are either specialized or computationally inefficient. We pr…

Cited by 11 Source ↗PDF ↗Scholar ↗