Notice: This website is continuously updated as new conference proceedings and journal volumes are published. Consequently, the figures presented here may differ from those reported in our publications.
A brief overview of our methodology is provided on the Methods page, along with references to publications documenting our approach and links to the code and data used in each study.
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.
Adaptive Data Debiasing through Bounded Exploration
Authors: Yifan Yang, Yang Liu, Parinaz Naghizadeh
NeurIPS 2022 | Venue PDF | LLM Run Details | Input Tokens: 19,837 Total number of tokens sent to the LLM as input for this paper's analysis. | Output Tokens: 3,775 Total number of tokens produced by the LLM (including reasoning/thinking tokens) for this paper's analysis.
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We illustrate the performance of our algorithm using experiments on synthetic and real-world datasets. |
| Researcher Affiliation | Academia | Yifan Yang Ohio State University EMAIL Yang Liu University of California, Santa Cruz EMAIL Parinaz Naghizadeh Ohio State University EMAIL |
| Pseudocode | Yes | Our active debiasing algorithm is summarized below. A pseudo-code is given in Appendix C. |
| Open Source Code | Yes | Our code is available at: https://github.com/Yifankevin/adaptive_data_debiasing. |
| Open Datasets | Yes | Adult dataset [12] and the FICO credit score dataset [37] pre-processed by [16]. |
| Dataset Splits | No | We use 2.5% of the data to obtain a biased estimate of the parameter . The remaining data arrives sequentially. |
| Hardware Specification | No | All experiments were run on a local computer. |
| Software Dependencies | No | No specific software versions (e.g., PyTorch 1.9, Python 3.8) are mentioned. |
| Experiment Setup | Yes | Our algorithm sets 1 = 50 and 0 = 60 percentiles, and exploration frequencies t are selected adaptively by both our algorithm and pure exploration. |