STEEL samples a finite set of PEFT adapters from a diffusion model and selects the best on the downstream support set, producing non-vacuous PAC-Bayes generalization certificates for low-shot LLM and vision transfer learning.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Model Diffusion for Certifiable Few-shot Transfer Learning
STEEL samples a finite set of PEFT adapters from a diffusion model and selects the best on the downstream support set, producing non-vacuous PAC-Bayes generalization certificates for low-shot LLM and vision transfer learning.