Unifying the Stochastic and the Adversarial Bandits with Knapsack

Authors: Anshuka Rangi, Massimo Franceschetti, Long Tran-Thanh

IJCAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical This work investigates the adversarial Bandits with Knapsack (Bw K) learning problem... We propose a novel algorithm EXP3.Bw K and show that the expected regret of the algorithm is order optimal in the budget. We then propose another algorithm EXP3++.Bw K... These results are then extended to a more general online learning setting, by designing another algorithm EXP3++.Lw K and providing its performance guarantees. Finally, we investigate the scenario where the costs of the actions are large and comparable to the budget. We show that for the adversarial setting, the achievable regret bounds scale at least linearly with the maximum cost for any learning algorithm...
Researcher Affiliation Academia Anshuka Rangi 1 , Massimo Franceschetti1 and Long Tran-Thanh 2 1 University of California, San Diego, 2 University of Southampton arangi@ucsd.edu, massimo@ece.ucsd.edu, L.Tran-Thanh@soton.ac.uk
Pseudocode Yes Algorithm 1 EXP3.Bw K
Open Source Code No The paper does not provide an explicit statement or link for open-source code related to the methodology described.
Open Datasets No The paper is theoretical and does not conduct empirical studies involving datasets, thus no information about public datasets is provided.
Dataset Splits No The paper is theoretical and does not conduct empirical studies, thus no information about training/validation/test splits is provided.
Hardware Specification No The paper is theoretical and does not describe any experiments, thus no hardware specifications are provided.
Software Dependencies No The paper describes algorithms theoretically and does not mention specific software dependencies or their versions.
Experiment Setup No The paper is theoretical and does not include an experimental evaluation section, thus no experimental setup details or hyperparameters are provided.