WST: Weakly Supervised Transducer for Automatic Speech Recognition

TL;DR

This paper introduces WST, a weakly supervised transducer model for speech recognition that is robust to high transcription error rates and reduces reliance on costly annotated data, outperforming existing methods.

Contribution

The paper presents a novel WST model that handles transcription errors without extra confidence estimation, improving weakly supervised ASR performance.

Findings

WST maintains performance with up to 70% transcription errors.

WST outperforms BTC and OTC in robustness and accuracy.

Implementation will be publicly available.

Abstract

The Recurrent Neural Network-Transducer (RNN-T) is widely adopted in end-to-end (E2E) automatic speech recognition (ASR) tasks but depends heavily on large-scale, high-quality annotated data, which are often costly and difficult to obtain. To mitigate this reliance, we propose a Weakly Supervised Transducer (WST), which integrates a flexible training graph designed to robustly handle errors in the transcripts without requiring additional confidence estimation or auxiliary pre-trained models. Empirical evaluations on synthetic and industrial datasets reveal that WST effectively maintains performance even with transcription error rates of up to 70%, consistently outperforming existing Connectionist Temporal Classification (CTC)-based weakly supervised approaches, such as Bypass Temporal Classification (BTC) and Omni-Temporal Classification (OTC). These results demonstrate the practical utility and robustness of WST in realistic ASR settings. The implementation will be publicly available.