{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/attention-based-end-to-end-models-for-small","title":"Attention-based End-to-End Models for Small-Footprint Keyword Spotting","arxiv_id":"1803.10916","date":"2018-03-29","proceeding":null,"authors":["Changhao Shan","Junbo Zhang","Yujun Wang","Lei Xie"],"abstract":"In this paper, we propose an attention-based end-to-end neural approach for\nsmall-footprint keyword spotting (KWS), which aims to simplify the pipelines of\nbuilding a production-quality KWS system. Our model consists of an encoder and\nan attention mechanism. The encoder transforms the input signal into a high\nlevel representation using RNNs. Then the attention mechanism weights the\nencoder features and generates a fixed-length vector. Finally, by linear\ntransformation and softmax function, the vector becomes a score used for\nkeyword detection. We also evaluate the performance of different encoder\narchitectures, including LSTM, GRU and CRNN. Experiments on real-world wake-up\ndata show that our approach outperforms the recent Deep KWS approach by a large\nmargin and the best performance is achieved by CRNN. To be more specific, with\n~84K parameters, our attention-based model achieves 1.02% false rejection rate\n(FRR) at 1.0 false alarm (FA) per hour.","url_abs":"http://arxiv.org/abs/1803.10916v1","url_pdf":"http://arxiv.org/pdf/1803.10916v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"attention-based-end-to-end-models-for-small","repo_url":"https://github.com/Kirili4ik/kws-attention-pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"attention-based-end-to-end-models-for-small","repo_url":"https://github.com/isadrtdinov/kws-attention","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"attention-based-end-to-end-models-for-small","repo_url":"https://github.com/sergevkim/KeywordSpotting","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"keyword-spotting","task_name":"Keyword Spotting"},{"task_slug":"small-footprint-keyword-spotting","task_name":"Small-Footprint Keyword Spotting"}],"methods":[{"method_slug":"gru","method_name":"GRU"},{"method_slug":"lstm","method_name":"LSTM"},{"method_slug":"sigmoid-activation","method_name":"Sigmoid Activation"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"tanh-activation","method_name":"Tanh Activation"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1803.10916","atlas_url":"https://app.syntology.ai/?focus=1803.10916","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}