Anatomy of Industrial Scale Multilingual ASR

TL;DR

This paper presents AssemblyAI's large-scale multilingual ASR system, highlighting its architecture, training data, and performance advantages over existing models in real-world, industrial applications.

Contribution

It introduces a novel industrial-scale ASR system with a unique architecture and training regimen, optimized for speed, noise robustness, and multilingual capabilities.

Findings

Competitive WERs against larger models

5x inference speedup over Whisper baseline

30% reduction in hallucination rate

Abstract

This paper describes AssemblyAI's industrial-scale automatic speech recognition (ASR) system, designed to meet the requirements of large-scale, multilingual ASR serving various application needs. Our system leverages a diverse training dataset comprising unsupervised (12.5M hours), supervised (188k hours), and pseudo-labeled (1.6M hours) data across four languages. We provide a detailed description of our model architecture, consisting of a full-context 600M-parameter Conformer encoder pre-trained with BEST-RQ and an RNN-T decoder fine-tuned jointly with the encoder. Our extensive evaluation demonstrates competitive word error rates (WERs) against larger and more computationally expensive models, such as Whisper large and Canary-1B. Furthermore, our architectural choices yield several key advantages, including an improved code-switching capability, a 5x inference speedup compared to an optimized Whisper baseline, a 30% reduction in hallucination rate on speech data, and a 90% reduction in ambient noise compared to Whisper, along with significantly improved time-stamp accuracy. Throughout this work, we adopt a system-centric approach to analyzing various aspects of fully-fledged ASR models to gain practically relevant insights useful for real-world services operating at scale.