Research map / Learning theory and algorithms / Optimal transport
An Optimal Transport Approach for Computing Adversarial Training Lower Bounds in Multiclass Classification
Nicolas Garcia Trillos, Matt Jacobs, Jakwang Kim, Matthew Werenski
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| truncate Sinkhorn approach to order 3 instead of full order (ablation) | Sinkhorn adversarial-risk lower bound · MNIST/CIFAR-10 (10 classes, 100 points each) | worst-case computational complexity: O~(100^10) -> O~(100^3) | Improved |
| increase truncation order from 2 to 3 (ablation) | adversarial risk lower bound · MNIST/CIFAR-10 (adversarial risk below 0.3) | adversarial risk | No clear effect |
| truncate to order 2 or 3 instead of full order 4 (ablation) | adversarial risk lower bound · MNIST/CIFAR-10 (4 classes, 50 points) | adversarial risk | No clear effect |
| use entropic Sinkhorn instead of exact LP | adversarial risk computation · MNIST/CIFAR-10 (10 classes, 100 points each) | computation time and risk value | Mixed |
| set entropic parameter eta to 0.1 instead of larger values (ablation) | Sinkhorn approach · Synthetic 6 classes, 2D, order 3 | adversarial risk | Improved |
Closest papers by meaning
- Lower Bounds on Adversarial Robustness from Optimal Transport (NeurIPS 2019 · similarity 0.93)
- Adversarial Risk via Optimal Transport and Optimal Couplings (ICML 2020 · similarity 0.92)
- Achieving Robustness in Classification Using Optimal Transport With Hinge Regularization (CVPR 2021 · similarity 0.91)
- The multimarginal optimal transport formulation of adversarial multiclass classification (ICML 2023 · similarity 0.90)
- Theoretical Analysis of Adversarial Learning: A Minimax Approach (NeurIPS 2019 · similarity 0.89)
- Analyzing and Improving Optimal-Transport-based Adversarial Networks (ICLR 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
