Interaction Asymmetry: A General Principle for Learning Composable Abstractions

Documenti analoghi

• Generation is Required for Data-Efficient Perception
  di: Brady, Jack, et al.
  Pubblicazione: (2025)
• From Pixels to Components: Eigenvector Masking for Visual Representation Learning
  di: Bizeul, Alice, et al.
  Pubblicazione: (2025)
• LAION-C: An Out-of-Distribution Benchmark for Web-Scale Vision Models
  di: Li, Fanfei, et al.
  Pubblicazione: (2025)
• Does CLIP's Generalization Performance Mainly Stem from High Train-Test Similarity?
  di: Mayilvahanan, Prasanna, et al.
  Pubblicazione: (2023)
• A Sparsity Principle for Partially Observable Causal Representation Learning
  di: Xu, Danru, et al.
  Pubblicazione: (2024)