{"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/efficient-keyword-spotting-using-dilated","title":"Efficient keyword spotting using dilated convolutions and gating","arxiv_id":"1811.07684","date":"2018-11-19","proceeding":null,"authors":["Alice Coucke","Mohammed Chlieh","Thibault Gisselbrecht","David Leroy","Mathieu Poumeyrol","Thibaut Lavril"],"abstract":"We explore the application of end-to-end stateless temporal modeling to\nsmall-footprint keyword spotting as opposed to recurrent networks that model\nlong-term temporal dependencies using internal states. We propose a model\ninspired by the recent success of dilated convolutions in sequence modeling\napplications, allowing to train deeper architectures in resource-constrained\nconfigurations. Gated activations and residual connections are also added,\nfollowing a similar configuration to WaveNet. In addition, we apply a custom\ntarget labeling that back-propagates loss from specific frames of interest,\ntherefore yielding higher accuracy and only requiring to detect the end of the\nkeyword. Our experimental results show that our model outperforms a max-pooling\nloss trained recurrent neural network using LSTM cells, with a significant\ndecrease in false rejection rate. The underlying dataset - \"Hey Snips\"\nutterances recorded by over 2.2K different speakers - has been made publicly\navailable to establish an open reference for wake-word detection.","url_abs":"http://arxiv.org/abs/1811.07684v2","url_pdf":"http://arxiv.org/pdf/1811.07684v2.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":"efficient-keyword-spotting-using-dilated","repo_url":"https://github.com/snipsco/keyword-spotting-research-datasets","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"NOASSERTION"}},{"paper_slug":"efficient-keyword-spotting-using-dilated","repo_url":"https://github.com/snipsco/tract","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"efficient-keyword-spotting-using-dilated","repo_url":"https://github.com/sonos/keyword-spotting-research-datasets","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"NOASSERTION"}},{"paper_slug":"efficient-keyword-spotting-using-dilated","repo_url":"https://github.com/sonos/tract","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"efficient-keyword-spotting-using-dilated","repo_url":"https://github.com/danFromTelAviv/key_words_spotting","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"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":"dilated-causal-convolution","method_name":"Dilated Causal Convolution"},{"method_slug":"lstm","method_name":"LSTM"},{"method_slug":"mixture-of-logistic-distributions","method_name":"Mixture of Logistic Distributions"},{"method_slug":"sigmoid-activation","method_name":"Sigmoid Activation"},{"method_slug":"tanh-activation","method_name":"Tanh Activation"},{"method_slug":"wavenet","method_name":"WaveNet"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1811.07684","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}