Skip to search formSkip to main contentSkip to account menu
  • Publications
  • Influence
Bandit Algorithms
sets of environments and policies respectively and ` : E ×Π→ [0, 1] a bounded loss function. Given a policy π let `(π) = (`(ν1, π), . . . , `(νN , π)) be the loss vector resulting from policy π.…
View via Publisher
Unifying PAC and Regret: Uniform PAC Bounds for Episodic Reinforcement Learning
TLDR
A new framework for theoretically measuring the performance of reinforcement learning algorithms called Uniform-PAC, which is a strengthening of the classical Probably Approximately Correct (PAC) framework, and simultaneously achieves optimal regret and PAC guarantees except for a factor of the horizon.
View PDF on arXiv
Causal Bandits: Learning Good Interventions via Causal Inference
TLDR
A new algorithm is proposed that exploits the causal feedback and proves a bound on its simple regret that is strictly better (in all quantities) than algorithms that do not use the additional causal information.
View PDF on arXiv
The End of Optimism? An Asymptotic Analysis of Finite-Armed Linear Bandits
TLDR
Surprisingly, the results show that no algorithm based on optimism or Thompson sampling will ever achieve the optimal rate, and indeed, can be arbitrarily far from optimal, even in very simple cases.
View PDF on arXiv
PAC Bounds for Discounted MDPs
TLDR
A new bound is proved for a modified version of Upper Confidence Reinforcement Learning (ucrl) with only cubic dependence on the horizon, which is unimprovable in all parameters except the size of the state/action space, where it depends linearly on the number of non-zero transition probabilities.
View PDF on arXiv
Learning with Good Feature Representations in Bandits and in RL with a Generative Model
TLDR
The Kiefer-Wolfowitz theorem is used to prove a positive result that by checking only a few actions, a learner can always find an action that is suboptimal with an error of at most $O(\epsilon \sqrt{d})$.
View PDF on arXiv
Conservative Bandits
We study a novel multi-armed bandit problem that models the challenge faced by a company wishing to explore new strategies to maximize revenue whilst simultaneously maintaining their revenue above a…
View PDF on arXiv
Degenerate Feedback Loops in Recommender Systems
TLDR
A novel theoretical analysis is provided that examines both the role of user dynamics and the behavior of recommender systems, disentangling the echo chamber from the filter bubble effect, and offers practical solutions to slow down system degeneracy.
View on ACM
Model Selection in Contextual Stochastic Bandit Problems
TLDR
A lower bound is shown that even when one of the base algorithms has $O(\log T)$ regret, in general it is impossible to get better than $\Omega(\sqrt{T})$ regret in model selection, even asymptotically.
View PDF on arXiv
TopRank: A practical algorithm for online stochastic ranking
TLDR
This work proposes a generalized click model that encompasses many existing models, including the position-based and cascade models, and motivates a novel online learning algorithm based on topological sort, which is called TopRank.
View PDF on arXiv
...
...
By clicking accept or continuing to use the site, you agree to the terms outlined in our Privacy Policy, Terms of Service, and Dataset License