No-Regret Learning in Unknown Games with Correlated Payoffs

Authors: Pier Giuseppe Sessa, Ilija Bogunovic, Maryam Kamgarpour, Andreas Krause

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

Reproducibility Variable Result LLM Response
Research Type Experimental We experimentally demonstrate the effectiveness of GP-MW in random matrix games, as well as real-world problems of traffic routing and movie recommendation. In our experiments, GP-MW consistently outperforms several baselines, while its performance is often comparable to methods that have access to full information feedback.
Researcher Affiliation Academia Pier Giuseppe Sessa ETH Zürich sessap@ethz.ch Ilija Bogunovic ETH Zürich ilijab@ethz.ch Maryam Kamgarpour ETH Zürich maryamk@ethz.ch Andreas Krause ETH Zürich krausea@ethz.ch
Pseudocode Yes Algorithm 1 The GP-MW algorithm for player i
Open Source Code No The paper does not provide an explicit statement or link for the open-source code of the described methodology.
Open Datasets Yes We consider the Sioux-Falls road network [14, 1], a standard benchmark model in the transportation literature. Transportation network test problems. http://www.bgu.ac.il/ bargera/tntp/. We use the Movie Lens-100K dataset [12] which provides a matrix of ratings for 1682 movies rated by 943 users.
Dataset Splits No The paper describes the datasets used (random matrix games, Sioux-Falls road network, Movie Lens-100K) but does not provide explicit train/validation/test split percentages, sample counts, or cross-validation methodology.
Hardware Specification No The paper does not provide specific details about the hardware used to run the experiments, such as CPU or GPU models, or memory specifications.
Software Dependencies No The paper does not provide specific software dependencies with version numbers used for its implementation or experiments.
Experiment Setup Yes We select K = 30 and generate 10 random matrices with r1 = r2 GP(0, k( , )), where k = k SE with l = 6. We set the noise to i t N(0, 1), and use T = 200. We use the same set of algorithm parameters as in [8]. For every agent i to run GP-MW, we use a composite kernel ki such that for every a1, a2 2 A, ki((ai ), (a i 2 )) , where ki 1 is a linear kernel and ki 2 is a polynomial kernel of degree n 2 {2, 4, 6}. T = 100 game rounds.