Research map / Label-efficient and robust learning / Recognition and long-tail
Mitigating Information Leakage in Image Representations: A Maximum Entropy Approach
Proteek Chandan Roy, Vishnu Naresh Boddeti
Check 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 |
|---|---|---|---|
| Replace minimizing sensitive-label likelihood with maximizing discriminator entropy | Adversarial representation learning · Extended Yale B | lighting accuracy: 57 -> 40 | Improved |
| Replace minimizing sensitive-label likelihood with maximizing discriminator entropy | Adversarial representation learning · Extended Yale B | identity accuracy: 89 -> 89 | No clear effect |
| Replace minimizing sensitive-label likelihood with maximizing discriminator entropy | Adversarial representation learning · German credit | sensitive attribute accuracy: 80.2% -> 72.7% | Improved |
| Replace minimizing sensitive-label likelihood with maximizing discriminator entropy | Adversarial representation learning · German credit | target accuracy: 74.4% -> 86.33% | Improved |
| Replace minimizing sensitive-label likelihood with maximizing discriminator entropy | Adversarial representation learning · Adult income | sensitive attribute accuracy: 67.7% -> 65.5% | Improved |
| Replace minimizing sensitive-label likelihood with maximizing discriminator entropy | Adversarial representation learning · Adult income | target accuracy: 84.4% -> 84.6% | No clear effect |
| Replace minimizing sensitive-label likelihood with maximizing discriminator entropy | Adversarial representation learning · Mixture of four Gaussians | adversary accuracy: 70% -> 63% | Improved |
| Replace minimizing sensitive-label likelihood with maximizing discriminator entropy | Adversarial representation learning · CIFAR-100 superclass/fine class | adversary entropy: 2.27 -> 2.27; 4.60 -> 4.60 | Mixed |
Closest papers by meaning
- Adversarially Learned Representations for Information Obfuscation and Inference (ICML 2019 · similarity 0.87)
- Learning Robust and Privacy-Preserving Representations via Information Theory (AAAI 2025 · similarity 0.86)
- Adversarial Image Perturbation for Privacy Protection -- A Game Theory Perspective (ICCV 2017 · similarity 0.86)
- Reducing information dependency does not cause training data privacy. Adversarially non-robust features do. (ICLR 2026 · similarity 0.86)
- PID: Prompt-Independent Data Protection Against Latent Diffusion Models (ICML 2024 · similarity 0.84)
- The Fundamental Limits of Least-Privilege Learning (ICML 2024 · similarity 0.84)
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.
