Research map / Label-efficient and robust learning / Recognition and long-tail
Lossy Common Information in a Learnable Gray-Wyner Network
Anderson de Andrade, Alon Harell, Ivan V. Bajic
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 the learnable Gray-Wyner three-channel codec instead of independent single-task coding | two-task vision codecs · three computer vision experiments | transmit-rate BD-rate: -81.58% | Improved |
| use Shared analysis transforms instead of Separated or Combined architectures (ablation) | learnable Gray-Wyner Network · synthetic two-task regression | rate-distortion | Improved |
| set beta to 3/2 rather than 1 or 2 (ablation) | transmit-receive objective · synthetic two-task regression | transmit and receive rate | Improved |
| use a mixture colour PMF with partly separable common information instead of dependent or independent PMFs | learnable Gray-Wyner Network · colored MNIST | rate-distortion | Hurt |
| add the common channel (Gray-Wyner) versus Independent coding on Mixture PMF (ablation) | learnable Gray-Wyner Network · colored MNIST Mixture PMF | transmit rate | Improved |
Closest papers by meaning
- Cross-Stitch Networks for Multi-Task Learning (CVPR 2016 · similarity 0.85)
- Taskology: Utilizing Task Relations at Scale (CVPR 2021 · similarity 0.85)
- Selective Amnesia using Contrastive Subnet Erasure for Class Level Unlearning in Vision Models (CVPR 2026 · similarity 0.85)
- Towards flexible perception with visual memory (ICML 2025 · similarity 0.84)
- Large Vision-Language Models Get Lost in Attention (ICML 2026 · similarity 0.84)
- A Comprehensive Information-Decomposition Analysis of Large Vision-Language Models (ICLR 2026 · 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.
