SynesLM: A Unified Approach for Audio-visual Speech Recognition and Translation via Language Model and Synthetic Data

TL;DR

SynesLM is a unified multimodal model that performs audio-visual speech recognition and translation tasks, leveraging general visual cues and synthetic data to achieve state-of-the-art results across multiple benchmarks.

Contribution

This work introduces SynesLM, a novel unified model that integrates multiple multimodal language tasks using general visual information and synthetic data augmentation.

Findings

Achieved state-of-the-art zero-shot AV-ASR performance with lower WER.

Outperformed previous models in visual-aided speech translation and translation tasks.

Demonstrated competitive performance on the How2 dataset.

Abstract

In this work, we present SynesLM, an unified model which can perform three multimodal language understanding tasks: audio-visual automatic speech recognition(AV-ASR) and visual-aided speech/machine translation(VST/VMT). Unlike previous research that focused on lip motion as visual cues for speech signals, our work explores more general visual information within entire frames, such as objects and actions. Additionally, we use synthetic image data to enhance the correlation between image and speech data. We benchmark SynesLM against the How2 dataset, demonstrating performance on par with state-of-the-art (SOTA) models dedicated to AV-ASR while maintaining our multitasking framework. Remarkably, for zero-shot AV-ASR, SynesLM achieved SOTA performance by lowering the Word Error Rate (WER) from 43.4% to 39.4% on the VisSpeech Dataset. Furthermore, our results in VST and VMT outperform the previous results, improving the BLEU score to 43.5 from 37.2 for VST, and to 54.8 from 54.4 for VMT.