Endpoint Detection for Streaming End-to-End Multi-talker ASR

TL;DR

This paper enhances streaming multi-talker speech recognition by integrating endpoint detection into the SURT model using an end-of-sentence token and latency penalties, achieving accurate detection with minimal accuracy loss.

Contribution

It introduces endpoint detection into the SURT framework with an end-of-sentence token and latency penalty, a novel approach for multi-talker end-to-end models.

Findings

Effective endpoint detection achieved without significant accuracy loss.

Latency penalty reduces detection delay substantially.

Model performs well on LibrispeechMix dataset.

Abstract

Streaming end-to-end multi-talker speech recognition aims at transcribing the overlapped speech from conversations or meetings with an all-neural model in a streaming fashion, which is fundamentally different from a modular-based approach that usually cascades the speech separation and the speech recognition models trained independently. Previously, we proposed the Streaming Unmixing and Recognition Transducer (SURT) model based on recurrent neural network transducer (RNN-T) for this problem and presented promising results. However, for real applications, the speech recognition system is also required to determine the timestamp when a speaker finishes speaking for prompt system response. This problem, known as endpoint (EP) detection, has not been studied previously for multi-talker end-to-end models. In this work, we address the EP detection problem in the SURT framework by introducing an end-of-sentence token as an output unit, following the practice of single-talker end-to-end models. Furthermore, we also present a latency penalty approach that can significantly cut down the EP detection latency. Our experimental results based on the 2-speaker LibrispeechMix dataset show that the SURT model can achieve promising EP detection without significantly degradation of the recognition accuracy.