TokenVerse: Towards Unifying Speech and NLP Tasks via Transducer-based ASR

TL;DR

TokenVerse introduces a unified transducer-based model that integrates multiple speech and NLP tasks into a single system, improving efficiency and performance over traditional cascaded pipelines.

Contribution

This paper presents the first unified transducer model that handles ASR, speaker change detection, endpointing, and NER simultaneously, streamlining conversational AI pipelines.

Findings

Up to 7.7% relative WER improvement.

Outperforms cascaded pipeline in individual tasks.

Effective on both public and private datasets.

Abstract

In traditional conversational intelligence from speech, a cascaded pipeline is used, involving tasks such as voice activity detection, diarization, transcription, and subsequent processing with different NLP models for tasks like semantic endpointing and named entity recognition (NER). Our paper introduces TokenVerse, a single Transducer-based model designed to handle multiple tasks. This is achieved by integrating task-specific tokens into the reference text during ASR model training, streamlining the inference and eliminating the need for separate NLP models. In addition to ASR, we conduct experiments on 3 different tasks: speaker change detection, endpointing, and NER. Our experiments on a public and a private dataset show that the proposed method improves ASR by up to 7.7% in relative WER while outperforming the cascaded pipeline approach in individual task performance. Our code is publicly available: https://github.com/idiap/tokenverse-unifying-speech-nlp