Adversarial Linear Contextual Bandits with Graph-Structured Side Observations
Authors: Lingda Wang, Bingcong Li, Huozhi Zhou, Georgios B. Giannakis, Lav R. Varshney, Zhizhen Zhao10156-10164
AAAI 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Numerical tests corroborate the efficiency of proposed algorithms. ... We conduct the numerical tests on synthetic data to demonstrate the efficiency of the novel EXP3-LGC-U and EXP3-LGC-IX algorithms. ... Figure 2 presents the expected cumulative regret2, where the results are averaged over 100 independent trials. |
| Researcher Affiliation | Academia | Lingda Wang1, Bingcong Li2, Huozhi Zhou1, Georgios B. Giannakis2, Lav R. Varshney1, Zhizhen Zhao1 1University of Illinois at Urbana-Champaign 2University of Minnesota Twin Cities {lingdaw2, hzhou35, varshney, zhizhenz}@illinois.edu, {lixx5599, georgios}@umn.edu |
| Pseudocode | Yes | Algorithm 1 EXP3-LGC-U Input: Learning rate η > 0, uniform exploration rate γ (0, 1), covariance Σ, and action set V . For t = 1, . . . , T, do: ... Algorithm 2 EXP3-LGC-IX Parameters: Learning rate ηt > 0, implicit exploration rate βt (0, 1), and covariance Σ,and action set V . For t = 1, . . . , T, do: |
| Open Source Code | No | The paper does not provide concrete access to source code for the methodology described. There are no explicit statements about releasing code or links to repositories. |
| Open Datasets | No | The paper uses synthetic data generated by the authors, providing details on its generation (e.g., "Each coordinate of context Xt (or Xt) is generated i.i.d. from the Bernoulli distribution with support {0, 1/d} and p = 0.5"). It does not provide access information for a publicly available or open dataset. |
| Dataset Splits | No | The paper describes generating synthetic data and simulating a bandit problem over a time horizon, but it does not specify explicit training/validation/test dataset splits. The evaluation is performed over the entire time horizon, characteristic of online learning problems, without traditional dataset partitioning. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., GPU/CPU models, memory amounts) used for running its experiments. It only mentions "numerical tests" on synthetic data. |
| Software Dependencies | No | The paper describes its algorithms and numerical tests but does not specify any ancillary software or library names with version numbers needed to replicate the experiment. |
| Experiment Setup | Yes | We consider a setting of K = 10 actions, with d = 10 dimensional contexts observed iteratively on a T = 10^5 time horizon. ... The parameters of EXP3-LGC-U and EXP3-LGC-IX are chosen according to Corollary 1 and Theorem 2, respectively. For EXP3-LGC-U and EXP3-LGC-IX , the parameter selection methods are identical as before, except for setting α(G) = K. The parameters of Robust Lin EXP3 are tuned exactly the same as those in Neu and Olkhovskaya (2020). |