Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].
Putting Gale & Shapley to Work: Guaranteeing Stability Through Learning
Authors: Hadi Hosseini, Sanjukta Roy, Duohan Zhang
NeurIPS 2024 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Finally, our empirical results demonstrate intriguing tradeoffs between stability and optimality of the proposed algorithms, further complementing our theoretical findings. Lastly, we validate our theoretical findings using empirical simulations (Section 6). |
| Researcher Affiliation | Academia | Hadi Hosseini Penn State University, USA EMAIL Sanjukta Roy University of Leeds, UK EMAIL Duohan Zhang* Penn State University, USA EMAIL |
| Pseudocode | Yes | Algorithm 1: Uniform sampling algorithm; Algorithm 2: Arm elimination algorithm; Algorithm 3: AE arm-DA algorithm |
| Open Source Code | No | The paper does not contain an explicit statement or link in the main body to open-source code for the described methodology. While the NeurIPS checklist mentions code in supplementary material, this information is not present in the provided paper text. |
| Open Datasets | No | we consider N = K = 20 and randomly generate preferences. In particular, we follow a similar experiment setting in Liu et al. [2021]: for each i, the true utilities {µi,1, µi,2, . . . , µi,20} are randomized permutations of the sequence {1, 2, . . . , 20} so that the minimum preference gap is fixed ( = 1) and algorithm performance exhibits relatively low variability. Arms preferences are generated the same way. |
| Dataset Splits | No | The paper describes generating synthetic data for simulations but does not specify explicit training, validation, or test dataset splits. |
| Hardware Specification | No | Experiments use bandit domain and algorithms can be run on a typical personal computer. Minimal compute resources are required to reproduce experiments in the paper. |
| Software Dependencies | No | The paper does not list specific software names with version numbers used for the experiments. |
| Experiment Setup | Yes | For this, we consider N = K = 20 and randomly generate preferences. In particular, we follow a similar experiment setting in Liu et al. [2021]: for each i, the true utilities {µi,1, µi,2, . . . , µi,20} are randomized permutations of the sequence {1, 2, . . . , 20} so that the minimum preference gap is fixed ( = 1) and algorithm performance exhibits relatively low variability. Arms preferences are generated the same way. We conduct 200 independent simulations, with each simulation featuring a randomized true preference profile. |