{"id":11652,"date":"2018-11-13T06:01:28","date_gmt":"2018-11-13T10:01:28","guid":{"rendered":"http:\/\/whamit.mit.edu\/?p=11652"},"modified":"2018-11-11T11:55:00","modified_gmt":"2018-11-11T15:55:00","slug":"complang-1115-hao-tang-mit-csail","status":"publish","type":"post","link":"https:\/\/ling-phil.mit.edu\/whamit\/2018\/11\/13\/complang-1115-hao-tang-mit-csail\/","title":{"rendered":"CompLang 11\/15 - Hao Tang (MIT CSAIL)"},"content":{"rendered":"<p><strong>Speaker:<\/strong>\u00a0Hao Tang\u00a0(MIT CSAIL)<\/p>\n<p><strong>Title:<\/strong>\u00a0Automatic speech recognition without linguistic knowledge?<\/p>\n<p><strong>Date and time:<\/strong>\u00a0Thursday, 11\/15, 5-6pm<\/p>\n<p><strong>Location:<\/strong>\u00a046-5165<\/p>\n<p><strong>Abstract<\/strong>:<\/p>\n<p>Building state-of-the-art speech recognizers, besides a large corpus of transcribed speech, require several additional ingredients, such as a phoneme inventory, a lexicon, and a language model.\u00a0 These ingredients carry linguistic constraints to make training more feasible and more sample efficient. Recently, there has been a push towards building a speech recognizer end to end, i.e., using few or even none of the aforementioned ingredients.\u00a0 This raises a fundamental question: is it possible to train a speech recognizer without any linguistic constraint?<\/p>\n<p>\u00a0\u00a0How much data do we need to make it possible?\u00a0 What linguistic constraints are necessary for building a speech recognizer?\u00a0 In this talk, I will review the inner workings of conventional and end-to-end speech recognizers, and to help answer some of those questions, I will present empirical results in training end-to-end speech recognizers without any linguistic constraints.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Speaker:\u00a0Hao Tang\u00a0(MIT CSAIL) Title:\u00a0Automatic speech recognition without linguistic knowledge? Date and time:\u00a0Thursday, 11\/15, 5-6pm Location:\u00a046-5165 Abstract: Building state-of-the-art speech recognizers, besides a large corpus of transcribed speech, require several additional ingredients, such as a phoneme inventory, a lexicon, and a language model.\u00a0 These ingredients carry linguistic constraints to make training more feasible and more sample [&hellip;]<\/p>\n","protected":false},"author":21,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-11652","post","type-post","status-publish","format-standard","hentry","category-announcements"],"_links":{"self":[{"href":"https:\/\/ling-phil.mit.edu\/whamit\/wp-json\/wp\/v2\/posts\/11652","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/ling-phil.mit.edu\/whamit\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/ling-phil.mit.edu\/whamit\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/ling-phil.mit.edu\/whamit\/wp-json\/wp\/v2\/users\/21"}],"replies":[{"embeddable":true,"href":"https:\/\/ling-phil.mit.edu\/whamit\/wp-json\/wp\/v2\/comments?post=11652"}],"version-history":[{"count":1,"href":"https:\/\/ling-phil.mit.edu\/whamit\/wp-json\/wp\/v2\/posts\/11652\/revisions"}],"predecessor-version":[{"id":11653,"href":"https:\/\/ling-phil.mit.edu\/whamit\/wp-json\/wp\/v2\/posts\/11652\/revisions\/11653"}],"wp:attachment":[{"href":"https:\/\/ling-phil.mit.edu\/whamit\/wp-json\/wp\/v2\/media?parent=11652"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/ling-phil.mit.edu\/whamit\/wp-json\/wp\/v2\/categories?post=11652"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/ling-phil.mit.edu\/whamit\/wp-json\/wp\/v2\/tags?post=11652"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}