Scalable Neural Network Compression and Pruning Using Hard Clustering and L1 Regularization
This addresses the problem of reducing model size and computational cost for machine learning practitioners, though it appears incremental as it builds on existing compression techniques.
The paper tackles neural network compression by proposing a method that combines hard clustering for quantization and L1 regularization for pruning, achieving results competitive with state-of-the-art methods while being simpler to implement.
We propose a simple and easy to implement neural network compression algorithm that achieves results competitive with more complicated state-of-the-art methods. The key idea is to modify the original optimization problem by adding K independent Gaussian priors (corresponding to the k-means objective) over the network parameters to achieve parameter quantization, as well as an L1 penalty to achieve pruning. Unlike many existing quantization-based methods, our method uses hard clustering assignments of network parameters, which adds minimal change or overhead to standard network training. We also demonstrate experimentally that tying neural network parameters provides less gain in generalization performance than changing network architecture and connectivity patterns entirely.