Human-Machine Interaction Speech Corpus from the ROBIN project

TL;DR

This paper presents a new Romanian speech corpus designed to enhance human-machine interaction in technical purchasing scenarios, including its creation, statistics, and impact on speech recognition and dialogue systems.

Contribution

It introduces the ROBINTASC corpus and evaluates its effectiveness in improving ASR and dialogue components for human-machine interaction.

Findings

Improved low-latency ASR performance with the corpus

Enhanced dialogue system responses using the corpus

Detailed corpus statistics and acquisition methodology

Abstract

This paper introduces a new Romanian speech corpus from the ROBIN project, called ROBIN Technical Acquisition Speech Corpus (ROBINTASC). Its main purpose was to improve the behaviour of a conversational agent, allowing human-machine interaction in the context of purchasing technical equipment. The paper contains a detailed description of the acquisition process, corpus statistics as well as an evaluation of the corpus influence on a low-latency ASR system as well as a dialogue component.