Oracle-Efficient Algorithms for Online Linear Optimization with Bandit Feedback

Authors: Shinji Ito, Daisuke Hatano, Hanna Sumita, Kei Takemura, Takuro Fukunaga, Naonori Kakimura, Ken-Ichi Kawarabayashi

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

Reproducibility Variable Result LLM Response
Research Type Theoretical Our contribution is to propose algorithms that offer optimal regret bounds of O( T) as well as low oracle complexity for both non-stochastic settings and stochastic settings. Our algorithm for non-stochastic settings has an oracle complexity of O(T) and is the first algorithm that achieves both a regret bound of O( T) and an oracle complexity of O(poly(T)), given only linear optimization oracles.
Researcher Affiliation Collaboration Shinji Ito NEC Corporation, The University of Tokyo i-shinji@nec.com Daisuke Hatano RIKEN AIP daisuke.hatano@riken.jp Hanna Sumita Tokyo Metropolitan University sumita@tmu.ac.jp Kei Takemura NEC Corporation kei_takemura@nec.com Takuro Fukunaga Chuo University, RIKEN AIP, JST PRESTO fukunaga.07s@g.chuo-u.ac.jp Naonori Kakimura Keio University kakimura@math.keio.ac.jp Ken-ichi Kawarabayashi National Institute of Informatics k-keniti@nii.ac.jp
Pseudocode Yes Algorithm 1 An oracle efficient algorithm for non-stochastic bandit linear optimization
Open Source Code No The paper does not contain any explicit statements about the release of open-source code for the described methodology, nor does it provide any links to a code repository.
Open Datasets No The paper presents theoretical algorithms and does not involve empirical training on specific datasets. Therefore, no information about publicly available training datasets is provided.
Dataset Splits No The paper focuses on theoretical contributions and does not describe empirical experiments that would require validation dataset splits. Therefore, no information regarding training/test/validation dataset splits is provided.
Hardware Specification No The paper focuses on theoretical algorithms and does not describe any empirical experiments or their computational execution. Therefore, no hardware specifications are mentioned.
Software Dependencies No The paper describes theoretical algorithms and does not mention any specific software dependencies or their version numbers required for replication.
Experiment Setup No The paper describes theoretical algorithms and their properties, not an empirical experiment. While Algorithm 1 lists theoretical parameters like 'Learning rate η > 0' and 'error bound ε > 0', these are not concrete experimental setup details like hyperparameters used in an actual computational run.