What is a Good Metric to Study Generalization of Minimax Learners?
Authors: Asuman Ozdaglar, Sarath Pattathil, Jiawei Zhang, Kaiqing Zhang
NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Theoretical | In this paper, we aim to answer this question by first showing that primal risk, a universal metric to study generalization in minimization, fails in simple examples of minimax problems. Furthermore, another popular metric, the primal-dual risk, also fails to characterize the generalization behavior for minimax problems with nonconvexity, due to non-existence of saddle points. We thus propose a new metric to study generalization of minimax learners: the primal gap, to circumvent these issues. Next, we derive generalization bounds for the primal gap in nonconvexconcave settings. |
| Researcher Affiliation | Academia | Massachusetts Institute of Technology University of Maryland, College Park |
| Pseudocode | Yes | These two algorithms are described in Algorithms 1 and 2 in Appendix E. |
| Open Source Code | No | The paper does not provide any concrete access to source code for the methodology described. It mentions "numerical verified in the literature" but does not offer its own implementation. |
| Open Datasets | No | The paper uses analytical and theoretical examples (e.g., 'Example 1 (Analytical example)', 'Example 2 (GAN-training example)'). While it references concepts from GANs and cites relevant papers, it does not provide access information for a specific public dataset used for empirical training. |
| Dataset Splits | No | The paper is theoretical and does not describe empirical experiments requiring training/validation/test splits. |
| Hardware Specification | No | The paper is theoretical and does not mention any hardware specifications used for experiments. |
| Software Dependencies | No | The paper is theoretical and does not mention specific software dependencies with version numbers. |
| Experiment Setup | No | The paper is theoretical and does not provide specific experimental setup details such as hyperparameters or training configurations. |