Induced Inflection-Set Keyword Search in Speech

TL;DR

This paper explores searching for lexeme sets in speech through inflectional variants, analyzing how search performance varies with inflection hypotheses and emphasizing the importance of curated paradigms and a new evaluation set.

Contribution

It introduces a method for lexeme-set search in speech, evaluates the impact of inflection hypotheses, and provides a new evaluation dataset for inflection generation performance.

Findings

Performance varies with the number of hypothesized inflections.

Curated inflectional paradigms improve search accuracy.

A new evaluation set aids in assessing inflection generation methods.

Abstract

We investigate the problem of searching for a lexeme-set in speech by searching for its inflectional variants. Experimental results indicate how lexeme-set search performance changes with the number of hypothesized inflections, while ablation experiments highlight the relative importance of different components in the lexeme-set search pipeline and the value of using curated inflectional paradigms. We provide a recipe and evaluation set for the community to use as an extrinsic measure of the performance of inflection generation approaches.