Multiobjective Lipschitz Bandits under Lexicographic Ordering
Authors: Bo Xue, Ji Cheng, Fei Liu, Yimu Wang, Qingfu Zhang
AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Numerical experiments confirm the effectiveness of our algorithm. In this section, we conduct numerical experiments to verify the effectiveness of our algorithms. |
| Researcher Affiliation | Academia | 1Department of Computer Science, City University of Hong Kong, Hong Kong, China 2The City University of Hong Kong Shenzhen Research Institute, Shenzhen, China 3Cheriton School of Computer Science, University of Waterloo, Waterloo, Canada |
| Pseudocode | Yes | Algorithm 1: Static Discretization under Lexicographic Ordering (SDLO) Algorithm 2: Multi-stage Decision-Making under Static Discretization (MSDM-SD) Algorithm 3: Adaptive Discretization under Lexicographic Ordering (ADLO) Algorithm 4: Multi-stage Decision-Making under Adaptive Discretization (MSDM-AD) |
| Open Source Code | No | The paper does not provide any explicit statement or link indicating the availability of open-source code for the described methodology. |
| Open Datasets | No | The paper describes a simulated experimental setup ("we set the arm space X = [0, 1] with a Euclidean metric on it" and "we set the payoff yi t = µi(xt)+ηt, where ηt is drawn from a Gaussian distribution with mean 0 and variance 1"), which involves generating data rather than using a pre-existing publicly available dataset. No concrete access information for a public dataset is provided. |
| Dataset Splits | No | The paper does not explicitly provide details about training, validation, or test dataset splits. It only mentions a 'time horizon T' for the experiments. |
| Hardware Specification | No | The paper does not provide any specific hardware details (e.g., exact GPU/CPU models, processor types, or memory amounts) used for running its experiments. |
| Software Dependencies | No | The paper does not provide specific ancillary software details with version numbers (e.g., library names like PyTorch 1.9 or solver names like CPLEX 12.4). |
| Experiment Setup | Yes | Following the existing experimental setup (Magureanu, Combes, and Proutiere 2014), we set the arm space X = [0, 1] with a Euclidean metric on it. The number of objectives is set as m = 3, and the expected payoff functions are given as µ1(x) = 1 minp {0.1,0.4,0.8} |x p|, µ2(x) = 1 2 minp {0.3,0.7} |x p| and µ3(x) = 1 2|x 0.3|. ... The time horizon T is 6 104, and thus the nearly optimal query parameter r for SDLO is 0.025, as stated in Corollary 1. The static arm set for SDLO and PF-LEX is constructed as A = {0.025 + 0.05 (k 1)|k [20]}. The confidence term (15) is scaled by a factor searched in [1e 2, 1]... |