ASTRA: Aligning Speech and Text Representations for Asr without Sampling

TL;DR

ASTRA introduces a new method for speech recognition that aligns speech and text representations without sampling, leveraging inherent model alignments to improve performance and simplify the process.

Contribution

It proposes a novel modality matching approach using a weighted RNNT objective that eliminates the need for sampling and duration prediction in ASR models.

Findings

Matches state-of-the-art performance on FLEURS benchmark

Avoids misalignment issues caused by upsampling

Simplifies training by removing duration prediction requirement

Abstract

This paper introduces ASTRA, a novel method for improving Automatic Speech Recognition (ASR) through text injection.Unlike prevailing techniques, ASTRA eliminates the need for sampling to match sequence lengths between speech and text modalities. Instead, it leverages the inherent alignments learned within CTC/RNNT models. This approach offers the following two advantages, namely, avoiding potential misalignment between speech and text features that could arise from upsampling and eliminating the need for models to accurately predict duration of sub-word tokens. This novel formulation of modality (length) matching as a weighted RNNT objective matches the performance of the state-of-the-art duration-based methods on the FLEURS benchmark, while opening up other avenues of research in speech processing.