Awesome Neural Network Reprogrammability

Welcome to the comprehensive documentation for neural network reprogrammability resources.

Overview

Neural network reprogrammability refers to methods that enable pre-trained neural networks to be adapted for new tasks with minimal parameter updates, leveraging existing learned representations.

  • Papers - Tabular view organized by reprogrammability taxonomy dimensions
  • Taxonomy - Classification system for reprogrammability methods
  • Evaluations - Benchmark results and comparisons

Quick Start

Browse the main README for the complete curated list of papers, datasets, and tools.