Evaluating Self-Supervised Speech Representations for Indigenous American Languages

TL;DR

This paper evaluates the effectiveness of self-supervised speech models on indigenous American languages, demonstrating strong performance and potential for low-resource language ASR applications.

Contribution

It introduces a new benchmark for indigenous languages and assesses large SSL models' performance on Quechua and others, highlighting their generalizability.

Findings

SSL models perform strongly on indigenous languages

Large-scale models show promise for low-resource ASR

Benchmark includes Quechua, Guarani, Bribri

Abstract

The application of self-supervision to speech representation learning has garnered significant interest in recent years, due to its scalability to large amounts of unlabeled data. However, much progress, both in terms of pre-training and downstream evaluation, has remained concentrated in monolingual models that only consider English. Few models consider other languages, and even fewer consider indigenous ones. In our submission to the New Language Track of the ASRU 2023 ML-SUPERB Challenge, we present an ASR corpus for Quechua, an indigenous South American Language. We benchmark the efficacy of large SSL models on Quechua, along with 6 other indigenous languages such as Guarani and Bribri, on low-resource ASR. Our results show surprisingly strong performance by state-of-the-art SSL models, showing the potential generalizability of large-scale models to real-world data.