Rethinking Incentives in Recommender Systems: Are Monotone Rewards Always Beneficial?
Authors: Fan Yao, Chuanhao Li, Karthik Abinav Sankararaman, Yiming Liao, Yan Zhu, Qifan Wang, Hongning Wang, Haifeng Xu
NeurIPS 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | The strength of BRM lies in three aspects: ... These merits of BRM are supported by our empirical studies, in which we developed simulated environments, demonstrating the welfare induced by BRM outperforms baseline mechanisms in M3. ... We conduct simulations on game instances ... We simulate the welfare curve produced by Algorithm 2 alongside five baseline mechanisms. |
| Researcher Affiliation | Collaboration | Fan Yao University of Virginia fy4bc@virginia.edu Chuanhao Li Yale University chuanhao.li.cl2637@yale.edu Karthik Abinav Sankararaman Meta karthikabinavs@meta.com Yiming Liao Meta yimingliao@meta.com Yan Zhu Google yanzhuyz@google.com Qifan Wang Meta wqfcr@meta.com Hongning Wang University of Virginia hw5x@virginia.edu Haifeng Xu University of Chicago haifengxu@uchicago.edu |
| Pseudocode | Yes | Algorithm 1 (sim Stra) Simulate content creators strategy evolving dynamic ... Algorithm 2 Optimize W in BRCM |
| Open Source Code | No | The paper does not provide a direct link or explicit statement about the release of its source code. |
| Open Datasets | Yes | Results on Movie Lens-1m dataset [11] are shown in Appendix 8.10. For the synthetic data, we consider a uniform distribution F on X as follows: we fix the embedding dimension d and randomly sample Y cluster centers... |
| Dataset Splits | Yes | we first performed a 5-fold cross-validation and obtain an averaged RMSE = 0.739 on the test sets, then train the user/item embeddings with the complete dataset. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., GPU/CPU models, memory) used for running its experiments. |
| Software Dependencies | No | The paper does not provide specific software dependencies with version numbers (e.g., library or solver names with version numbers). |
| Experiment Setup | Yes | The parameters are set to T = 1000, L1 = 200, L2 = 5, η1 = η2 = 0.1, f (0) = (1, 1, 1, 1, 1, 0, , 0). ... The parameters of Algorithm 2 are set to L1 = 100, L2 = 5, η1 = 0.5, η2 = 0.1, f (0) = (1, 1, 1, 1, 1, 0, , 0). |