Research map / Learning theory and algorithms / Kernels and regression theory
Bound by semanticity: universal laws governing the generalization-identification tradeoff
Marco Nurisso, Jesseba Fernando, Raj Deshpande, Alan Perotti, Raja Marjieh, Steven M Frankland, Richard Lewis, Taylor Whittington Webb, Declan Iain Campbell, Francesco Vaccarino, Jonathan D. Cohen, Giovanni Petri
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 |
|---|---|---|---|
| train the ReLU toy network on the similarity task instead of pure reconstruction (ablation) | toy ReLU network (l=50, m=10) · circle similarity task | p_S and p_I | Improved |
| increase generalization weight alpha in the loss (ablation) | ResNet-50 fine-tuned on bird species · bird species identification vs phylogenetic similarity | generalization vs identification accuracy | Mixed |
| increase probe distance from reference years | gemma-2b-it, Llama-3.2-3B-Instruct, Qwen2.5-7B-Instruct · year similarity task | accuracy | Hurt |
Closest papers by meaning
- Flatness is Necessary, Neural Collapse is Not: Rethinking Generalization via Grokking (NeurIPS 2025 · similarity 0.86)
- Scaling can lead to compositional generalization (NeurIPS 2025 · similarity 0.85)
- Position: The Platonic Representation Hypothesis (ICML 2024 · similarity 0.85)
- A mathematical theory of balancing relational generalization and memorization (NeurIPS 2026 · similarity 0.85)
- Minimal Semantic Sufficiency Meets Unsupervised Domain Generalization (NeurIPS 2025 · similarity 0.85)
- To Grok or not to Grok: Disentangling Generalization and Memorization on Corrupted Algorithmic Datasets (ICLR 2024 · 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.
