Geometric Limits of Knowledge Distillation: A Minimum-Width Theorem via Superposition Theory

Documenti analoghi

• Causal Dimensionality of Transformer Representations: Measurement, Scaling, and Layer Structure
  di: Sarkar, Nilesh, et al.
  Pubblicazione: (2026)
• Spectral Superposition: A Theory of Feature Geometry
  di: Ivanov, Georgi, et al.
  Pubblicazione: (2026)
• The Graphon Limit Hypothesis: Understanding Neural Network Pruning via Infinite Width Analysis
  di: Pham, Hoang, et al.
  Pubblicazione: (2025)
• Oh! We Freeze: Improving Quantized Knowledge Distillation via Signal Propagation Analysis for Large Language Models
  di: Bhardwaj, Kartikeya, et al.
  Pubblicazione: (2024)
• Class Incremental Fault Diagnosis under Limited Fault Data via Supervised Contrastive Knowledge Distillation
  di: Zhang, Hanrong, et al.
  Pubblicazione: (2025)