Research map / Label-efficient and robust learning / Recognition and long-tail
Learning Invariant Representations using Inverse Contrastive Loss
Aditya Kumar Akash, Vishnu Suresh Lokhande, Sathya N. Ravi, Vikas Singh
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 |
|---|---|---|---|
| Use ICL instead of the unregularized VAE | VAE · MNIST | adversarial invariance measure: 46 \pm 4 -> 32 \pm 0 | Improved |
| Use ICL instead of the strongest reconstruction baseline, CAI | VAE · MNIST | reconstruction error: 11.8 \pm 0.3 -> 16.6 \pm 0.1 | Hurt |
| Use ICL instead of MI | VIB · Adult | adversarial invariance measure: 78 \pm 2 -> 75 \pm 2 | Improved |
| Use ICL instead of the unregularized VIB | VIB · Adult | prediction accuracy: 84 \pm 0 -> 83 \pm 0 | Hurt |
| Use ICL instead of MMD_f | VIB · German | prediction accuracy: 74 \pm 1 -> 75 \pm 2 | Improved |
| Use ICL instead of MMD_{-s} | VIB · German | adversarial invariance measure: 75 \pm 2 -> 75 \pm 2 | No clear effect |
| Use ICL instead of UAI | VIB · MNIST-ROT | prediction accuracy: 98 -> 96 \pm 0 | Hurt |
| Use ICL instead of UAI | VIB · MNIST-ROT | adversarial invariance measure: 34 -> 33 \pm 1 | Improved |
| Use ICL instead of MMD_f | VIB · ADNI | adversarial invariance measure: 49 \pm 3 -> 46 \pm 7 | Improved |
| Use ICL instead of MMD_f | VIB · ADNI | prediction accuracy: 86 \pm 1 -> 84 \pm 3 | Hurt |
| Use ICL instead of CAI | ResNet-18 · Adult (Age) | adversarial invariance measure (MSE): 129 \pm 10 -> 161 \pm 15 | Improved |
| Use ICL instead of UAI | ResNet-18 · Adult (Age) | prediction accuracy: 84 \pm 0 -> 83 \pm 0 | Hurt |
Closest papers by meaning
- I-Con: A Unifying Framework for Representation Learning (ICLR 2025 · similarity 0.86)
- Symmetric Neural-Collapse Representations with Supervised Contrastive Loss: The Impact of ReLU and Batching (ICLR 2024 · similarity 0.85)
- Bridging Mini-Batch and Asymptotic Analysis in Contrastive Learning: From InfoNCE to Kernel-Based Losses (ICML 2024 · similarity 0.85)
- Engineering the Neural Collapse Geometry of Supervised-Contrastive Loss (Student Abstract) (AAAI 2024 · similarity 0.85)
- Decoupled Contrastive Learning for Long-Tailed Recognition (AAAI 2024 · similarity 0.85)
- On the Similarities of Embeddings in Contrastive Learning (ICML 2025 · similarity 0.85)
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.
