Compressing Neural Networks with the Hashing Trick

TL;DR

This paper introduces HashedNets, a neural network architecture that uses hashing to share weights among connections, significantly reducing model size while maintaining performance, enabling deployment on memory-limited devices.

Contribution

The paper proposes a novel hashing-based architecture for neural networks that drastically reduces model size without sacrificing accuracy.

Findings

HashedNets reduce model size substantially.

Weight sharing via hashing preserves generalization.

No additional memory overhead from hashing.

Abstract

As deep nets are increasingly used in applications suited for mobile devices, a fundamental dilemma becomes apparent: the trend in deep learning is to grow models to absorb ever-increasing data set sizes; however mobile devices are designed with very little memory and cannot store such large models. We present a novel network architecture, HashedNets, that exploits inherent redundancy in neural networks to achieve drastic reductions in model sizes. HashedNets uses a low-cost hash function to randomly group connection weights into hash buckets, and all connections within the same hash bucket share a single parameter value. These parameters are tuned to adjust to the HashedNets weight sharing architecture with standard backprop during training. Our hashing procedure introduces no additional memory overhead, and we demonstrate on several benchmark data sets that HashedNets shrink the storage requirements of neural networks substantially while mostly preserving generalization performance.