A Spectral Theory of Compositional Learning
cs.LG
Submitted: 2026-09-27
Updated: 2026-09-27
Terminology
Sources
- Convolutional 2D Knowledge Graph Embeddings
- An analytic theory of generalization dynamics and transfer learning in deep linear networks
- Progress measures for grokking via mechanistic interpretability
- Emergence of Hidden Capabilities: Exploring Learning Dynamics in Concept Space
- Exact solutions to the nonlinear dynamics of learning in deep linear neural networks
- RotatE: Knowledge Graph Embedding by Relational Rotation in Complex Space
- Compositional Generalization from First Principles
- Swing-by Dynamics in Concept Learning and Compositional Generalization
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks