Semantic enrichment towards efficient speech representations

TL;DR

This paper explores semantic enrichment of speech representations using the SAMU-XLSR model, focusing on improving spoken language understanding in low-resource languages and evaluating cross-domain capabilities.

Contribution

It introduces in-domain semantic specialization of SAMU-XLSR with limited data and assesses its effectiveness across different languages and domains.

Findings

Semantic enrichment improves SLU performance.

In-domain training enhances language-specific understanding.

Cross-domain capabilities are maintained with enriched models.

Abstract

Over the past few years, self-supervised learned speech representations have emerged as fruitful replacements for conventional surface representations when solving Spoken Language Understanding (SLU) tasks. Simultaneously, multilingual models trained on massive textual data were introduced to encode language agnostic semantics. Recently, the SAMU-XLSR approach introduced a way to make profit from such textual models to enrich multilingual speech representations with language agnostic semantics. By aiming for better semantic extraction on a challenging Spoken Language Understanding task and in consideration with computation costs, this study investigates a specific in-domain semantic enrichment of the SAMU-XLSR model by specializing it on a small amount of transcribed data from the downstream task. In addition, we show the benefits of the use of same-domain French and Italian benchmarks for low-resource language portability and explore cross-domain capacities of the enriched SAMU-XLSR.