PAC-Bayesian Lifelong Learning For Multi-Armed Bandits

Autor: Hamish Flynn, David Reeb, Melih Kandemir, Jan Peters
Jazyk: angličtina
Rok vydání: 2022
Předmět:
Zdroj: Flynn, H, Reeb, D, Kandemir, M & Peters, J 2022, ' PAC-Bayesian lifelong learning for multi-armed bandits ', Data Mining and Knowledge Discovery, vol. 36, pp. 841-876 . https://doi.org/10.1007/s10618-022-00825-4
DOI: 10.1007/s10618-022-00825-4
Popis: We present a PAC-Bayesian analysis of lifelong learning. In the lifelong learning problem, a sequence of learning tasks is observed one-at-a-time, and the goal is to transfer information acquired from previous tasks to new learning tasks. We consider the case when each learning task is a multi-armed bandit problem. We derive lower bounds on the expected average reward that would be obtained if a given multi-armed bandit algorithm was run in a new task with a particular prior and for a set number of steps. We propose lifelong learning algorithms that use our new bounds as learning objectives. Our proposed algorithms are evaluated in several lifelong multi-armed bandit problems and are found to perform better than a baseline method that does not use generalisation bounds.
29 pages, 5 figures
Databáze: OpenAIRE