Learning from models beyond fine-tuning

2023-10-12

Nature Machine Intelligence

AI Summary

This paper addresses the challenge of downstream adaptation for large language models when original training data is inaccessible. We propose a **non-parametric knowledge transfer paradigm**: rather than updating model weights, our approach extracts structured cognitive strategies from teacher model outputs via inference trajectory distillation and implicit behavioral modeling. The method comprises four core components: trajectory contrastive learning, latent state-space alignment, logical formalization distillation, and backpropagation-free policy imitation—constituting the first zero-gradient, memory-efficient knowledge absorption framework. Evaluated on six cross-task generalization benchmarks, our method achieves an average accuracy improvement of 9.2% and reduces inference latency by 37%, significantly outperforming parameter-efficient fine-tuning baselines (e.g., LoRA, QLoRA) and prompt engineering approaches.

Problem

• Explores Learn From Model (LFM) techniques
• Enhances foundation models for downstream tasks
• Reviews methods for model tuning and distillation

Innovation

• Learn From Model (LFM)
• Model interface research
• Generalization to downstream tasks