End-to-End Open Vocabulary Keyword Search With Multilingual Neural Representations

TL;DR

This paper introduces a multilingual neural model for open-vocabulary keyword search that simplifies the pipeline and outperforms traditional ASR-based systems on long and out-of-vocabulary queries.

Contribution

It extends previous neural ASR-free keyword search models with multilingual pretraining, enhancing performance and analysis of the approach.

Findings

Multilingual training improves model performance.

Outperforms ASR-based systems on long and out-of-vocabulary queries.

Maintains an efficient, simplified search pipeline.

Abstract

Conventional keyword search systems operate on automatic speech recognition (ASR) outputs, which causes them to have a complex indexing and search pipeline. This has led to interest in ASR-free approaches to simplify the search procedure. We recently proposed a neural ASR-free keyword search model which achieves competitive performance while maintaining an efficient and simplified pipeline, where queries and documents are encoded with a pair of recurrent neural network encoders and the encodings are combined with a dot-product. In this article, we extend this work with multilingual pretraining and detailed analysis of the model. Our experiments show that the proposed multilingual training significantly improves the model performance and that despite not matching a strong ASR-based conventional keyword search system for short queries and queries comprising in-vocabulary words, the proposed model outperforms the ASR-based system for long queries and queries that do not appear in the training data.