Decentralizing Feature Extraction with Quantum Convolutional Neural Network for Automatic Speech Recognition

TL;DR

This paper introduces a quantum convolutional neural network-based decentralized feature extraction method for speech recognition, enhancing privacy and achieving high accuracy in federated learning settings.

Contribution

It presents a novel quantum-based feature extraction framework in federated learning that improves privacy preservation and recognition accuracy.

Findings

Achieved 95.12% accuracy on Google Speech Commands Dataset

Demonstrated enhanced privacy protection in decentralized speech recognition

Provided insights into quantum circuit encoder architectures

Abstract

We propose a novel decentralized feature extraction approach in federated learning to address privacy-preservation issues for speech recognition. It is built upon a quantum convolutional neural network (QCNN) composed of a quantum circuit encoder for feature extraction, and a recurrent neural network (RNN) based end-to-end acoustic model (AM). To enhance model parameter protection in a decentralized architecture, an input speech is first up-streamed to a quantum computing server to extract Mel-spectrogram, and the corresponding convolutional features are encoded using a quantum circuit algorithm with random parameters. The encoded features are then down-streamed to the local RNN model for the final recognition. The proposed decentralized framework takes advantage of the quantum learning progress to secure models and to avoid privacy leakage attacks. Testing on the Google Speech Commands Dataset, the proposed QCNN encoder attains a competitive accuracy of 95.12% in a decentralized model, which is better than the previous architectures using centralized RNN models with convolutional features. We also conduct an in-depth study of different quantum circuit encoder architectures to provide insights into designing QCNN-based feature extractors. Neural saliency analyses demonstrate a correlation between the proposed QCNN features, class activation maps, and input spectrograms. We provide an implementation for future studies.