Research map / Probabilistic and scientific modeling / Molecules and proteins
Bridging Functional and Representational Similarity via Usable Information
Antonio Almudévar, Alfonso Ortega
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 bidirectional stitching instead of one stitcher | stitching-based functional similarity · 58 vision models, 12,459 layer pairs | accuracy-ratio difference between directions | Improved |
| widen predictive family from orthogonal to orthogonal+scale to affine (ablation) | representational similarity · 58 vision models | representational similarity | Improved |
| compare CKA with usable-information similarity | CKA · MLP layer pairs | R2 of fit: R2=0.937 | Improved |
| compare RSA with usable-information similarity | RSA · MLP layer pairs | R2 of fit: R2=0.912 | Improved |
| compare SVCCA with usable-information similarity | SVCCA · MLP layer pairs | R2 of fit: R2=0.650 | Mixed |
| use coarse 20-superclass task instead of 100 fine classes | functional similarity · CIFAR-100 | functional similarity | Improved |
| high representational similarity as condition for functional similarity | representational similarity · 58 vision models | P(accuracy ratio > 0.95) | Improved |
| low representational similarity as condition for functional similarity | representational similarity · 58 vision models | functional similarity | No clear effect |
Closest papers by meaning
- Functional Alignment Can Mislead: Examining Model Stitching (ICML 2025 · similarity 0.90)
- Objective drives the consistency of representational similarity across datasets (ICML 2025 · similarity 0.89)
- Not all solutions are created equal: An analytical dissociation of functional and representational similarity in deep linear neural networks (ICML 2025 · similarity 0.88)
- Towards a learning theory of representation alignment (ICLR 2025 · similarity 0.88)
- Latent Functional Maps: a spectral framework for representation alignment (NeurIPS 2024 · similarity 0.88)
- How Not to Stitch Representations to Measure Similarity: Task Loss Matching Versus Direct Matching (AAAI 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.
