CEG4N: Counter-Example Guided Neural Network Quantization Refinement

TL;DR

CEG4N is a novel method that refines neural network quantization by using counter-examples to ensure accuracy is maintained, resulting in models with significantly improved precision over existing techniques.

Contribution

This paper introduces CEG4N, a new approach combining search-based quantization and equivalence verification to improve neural network accuracy after quantization.

Findings

Achieves up to 72% better accuracy than state-of-the-art methods.

Successfully quantizes diverse neural networks without accuracy loss.

Demonstrates effectiveness across various benchmarks.

Abstract

Neural networks are essential components of learning-based software systems. However, their high compute, memory, and power requirements make using them in low resources domains challenging. For this reason, neural networks are often quantized before deployment. Existing quantization techniques tend to degrade the network accuracy. We propose Counter-Example Guided Neural Network Quantization Refinement (CEG4N). This technique combines search-based quantization and equivalence verification: the former minimizes the computational requirements, while the latter guarantees that the network's output does not change after quantization. We evaluate CEG4N~on a diverse set of benchmarks, including large and small networks. Our technique successfully quantizes the networks in our evaluation while producing models with up to 72% better accuracy than state-of-the-art techniques.