Adaptively Learning to Select-Rank in Online Platforms
Authors: Jingyuan Wang, Perry Dong, Ying Jin, Ruohan Zhan, Zhengyuan Zhou
ICML 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments conducted on both simulated and real-world datasets demonstrate our algorithm outperforms the baseline.We hereby provide empirical performances of UCR and G-MLE (the comparison of which is in Remark 3.1) on a simulated dataset and a real-world dataset, with the goal of illuminating how the two algorithms perform across different environments. |
| Researcher Affiliation | Collaboration | 1Stern School of Business, New York University 2EECS, UC Berkeley 3Arena Technologies 4Department of Statistics, Stanford University 5IEDA, Hong Kong University of Science and Technology (HKUST) 6HKUST Shenzhen-Hong Kong Collaborative Innovation Research Institute. |
| Pseudocode | Yes | Algorithm 1 Upper Confidence Ranking (UCR) and Algorithm 2 Subroutine: Upper Confidence Ranking via Maximum Weighted Bipartite Matching |
| Open Source Code | Yes | The python code for executing the experiment can be found in https://github.com/arena-tools/ranking-agent. |
| Open Datasets | No | The paper states "An offline dataset from the historical data... is used to learn a simulator." and "The dataset comprises 13,717 samples and 436 unique items, all used for simulator training." It also notes "For data-privacy reasons, the name of the company is not disclosed." and provides no public access link, DOI, or formal citation for this dataset. |
| Dataset Splits | No | The paper describes using a "simulator" trained on a real-world dataset but does not provide specific train/validation/test dataset splits, percentages, or sample counts, nor does it refer to predefined splits with citations. |
| Hardware Specification | No | The paper describes the experimental setup and results but does not specify any particular hardware components such as GPU or CPU models, memory, or cloud instance types used for running the experiments. |
| Software Dependencies | No | The paper mentions "The python code for executing the experiment" but does not provide specific version numbers for Python or any other software dependencies, libraries, or solvers used in the experiments. |
| Experiment Setup | Yes | For each setting we run several upper confidence parameters of UCR and present the regret curves of these instances along with those of the baseline G-MLE approach. All experiments have a initialization phase T0 = 5. |