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 Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The Shift Toward Open and Reproducible AI Research," Under Review, 2026. doi: 10.48550/arXiv.2606.16974..
Robust Policy Gradient against Strong Data Corruption
Authors: Xuezhou Zhang, Yiding Chen, Xiaojin Zhu, Wen Sun
ICML 2021 | Venue PDF | LLM Run Details | Input Tokens: 19,519 Total number of tokens sent to the LLM as input for this paper's analysis. | Output Tokens: 3,275 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 | Complimentary to the theoretical results, we show that a neural implementation of FPG achieves strong robust learning performance on the Mu Jo Co continuous control benchmarks. ... The experiment results are shown in Figure 1. |
| Researcher Affiliation | Academia | 1Department of Computer Sciences, University of Wisconsin Madison 2Cornell University. |
| Pseudocode | Yes | Algorithm 1 dπ ν sampler and Qπ estimator; Algorithm 2 Natural Policy Gradient (NPG); Algorithm 3 Robust Linear Regression via SEVER |
| Open Source Code | No | The paper does not provide a statement or link indicating that the source code for their methodology is publicly available. |
| Open Datasets | Yes | we show that a neural implementation of FPG achieves strong robust learning performance on the Mu Jo Co continuous control benchmarks (Todorov et al., 2012) |
| Dataset Splits | No | The paper mentions using MuJoCo benchmarks but does not specify details regarding training, validation, or test dataset splits. |
| Hardware Specification | No | The paper does not specify the hardware used for running the experiments. |
| Software Dependencies | No | The paper mentions software like TRPO, PyTorch, and SEVER but does not provide specific version numbers for any of them. |
| Experiment Setup | Yes | Throughout the experiment, we set the contamination level ε = 0.01, and tune δ among the values of [1, 2, 4, 8, 16, 32, 64]... All experiments are repeated with 3 random seeds and the mean and standard deviations are plotted in the figures. ... The pseudo-code and implementation details are discussed in appendix G. |