Sparsification via Compressed Sensing for Automatic Speech Recognition

TL;DR

This paper introduces a compressed sensing based pruning method for automatic speech recognition models, effectively reducing model size and latency while maintaining accuracy, outperforming existing pruning techniques.

Contribution

It proposes a novel CSP approach that integrates compressed sensing into model training for improved sparse pruning in ASR models.

Findings

CSP outperforms existing pruning methods in ASR tasks.

The approach reduces model size and latency effectively.

It maintains high accuracy despite aggressive sparsification.

Abstract

In order to achieve high accuracy for machine learning (ML) applications, it is essential to employ models with a large number of parameters. Certain applications, such as Automatic Speech Recognition (ASR), however, require real-time interactions with users, hence compelling the model to have as low latency as possible. Deploying large scale ML applications thus necessitates model quantization and compression, especially when running ML models on resource constrained devices. For example, by forcing some of the model weight values into zero, it is possible to apply zero-weight compression, which reduces both the model size and model reading time from the memory. In the literature, such methods are referred to as sparse pruning. The fundamental questions are when and which weights should be forced to zero, i.e. be pruned. In this work, we propose a compressed sensing based pruning (CSP) approach to effectively address those questions. By reformulating sparse pruning as a sparsity inducing and compression-error reduction dual problem, we introduce the classic compressed sensing process into the ML model training process. Using ASR task as an example, we show that CSP consistently outperforms existing approaches in the literature.