Query-by-example search often uses dynamic time warping (DTW) for comparing queries and proposed matching segments. Recent work has shown that comparing speech segments by representing them as fixed-dimensional vectors — acoustic word embeddings — and measuring their vector distance (e.g., cosine distance) can discriminate between words more accurately than DTW-based approaches. We consider an approach to query-by-example search that embeds both the query and database segments according to a neural model, followed by nearest-neighbor search to find the matching segments. Earlier work on embedding-based query-by-example, using template-based acoustic word embeddings, achieved competitive performance. We find that our embeddings, based on recurrent neural networks trained to optimize word discrimination, achieve substantial improvements in performance and run-time efficiency over the previous approaches.
Cite as: Settle, S., Levin, K., Kamper, H., Livescu, K. (2017) Query-by-Example Search with Discriminative Neural Acoustic Word Embeddings. Proc. Interspeech 2017, 2874-2878, doi: 10.21437/Interspeech.2017-1592
@inproceedings{settle17_interspeech,
author={Shane Settle and Keith Levin and Herman Kamper and Karen Livescu},
title={{Query-by-Example Search with Discriminative Neural Acoustic Word Embeddings}},
year=2017,
booktitle={Proc. Interspeech 2017},
pages={2874--2878},
doi={10.21437/Interspeech.2017-1592}
}