Improving Out-of-Distribution Generalization by Adversarial Training with Structured Priors
Authors: Qixun Wang, Yifei Wang, Hong Zhu, Yisen Wang
NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments on Domain Bed benchmark show that our proposed approaches outperform Empirical Risk Minimization (ERM) and sample-wise AT. Our code is available at https://github.com/NOVAglow646/NIPS22-MAT-and-LDAT-for-OOD. |
| Researcher Affiliation | Collaboration | 1 Key Lab. of Machine Perception (Mo E), School of Intelligence Science and Technology, Peking University 2 School of Mathematical Sciences, Peking University 3 Huawei Noah s Ark Lab 4 Institute for Artificial Intelligence, Peking University |
| Pseudocode | Yes | Algorithm 1 Detailed Training Procedure of MAT |
| Open Source Code | No | Did you include the code, data, and instructions needed to reproduce the main experimental results (either in the supplemental material or as a URL)? [No] It will release upon acceptance. |
| Open Datasets | Yes | We conduct experiments on the Domain Bed benchmark [9], a testbed for OOD generalization that implements consistent experimental protocols across various approaches to ensure fair comparisons. We evaluate on PACS [14], Office Home [15], VLCS [16], NICO [17], and Colored MNIST [1]. |
| Dataset Splits | Yes | For the split, we use the standard setup provided by Domain Bed, which uses a 75/25 train/test split, with 20% of the training data used as validation. |
| Hardware Specification | No | Did you include the total amount of compute and the type of resources used (e.g., type of GPUs, internal cluster, or cloud provider)? [No] |
| Software Dependencies | No | The paper mentions using ResNet-18 as a backbone and Grad Cam for visualization, but it does not specify version numbers for these or other software dependencies. |
| Experiment Setup | Yes | We run ERM and AT on four OOD datasets: PACS [14], Office Home [15], VLCS [16], and NICO [17] with a fixed set of hyperparameters (detailed experimental settings can be found in Appendix C.1). ... See Appendix C.2 for more details. ... The complete setup of the hyperparameters for MAT and LDAT is provided in Appendix C.2. |