InterSpeech 2021

Audio Retrieval with Natural Language Queries
(Oral presentation)

Andreea-Maria Oncescu (University of Oxford, UK), A. Sophia Koepke (Universität Tübingen, Germany), João F. Henriques (University of Oxford, UK), Zeynep Akata (Universität Tübingen, Germany), Samuel Albanie (University of Oxford, UK)
We consider the task of retrieving audio using free-form natural language queries. To study this problem, which has received limited attention in the existing literature, we introduce challenging new benchmarks for text-based audio retrieval using text annotations sourced from the AUDIOCAPS and CLOTHO datasets. We then employ these benchmarks to establish baselines for cross-modal audio retrieval, where we demonstrate the benefits of pre-training on diverse audio tasks. We hope that our benchmarks will inspire further research into cross-modal text-based audio retrieval with free-form text queries.