{"id":183106,"date":"2007-02-13T00:00:00","date_gmt":"2009-10-31T10:19:12","guid":{"rendered":"https:\/\/www.microsoft.com\/en-us\/research\/msr-research-item\/bayesian-inference-of-grammars\/"},"modified":"2016-09-09T09:43:25","modified_gmt":"2016-09-09T16:43:25","slug":"bayesian-inference-of-grammars","status":"publish","type":"msr-video","link":"https:\/\/www.microsoft.com\/en-us\/research\/video\/bayesian-inference-of-grammars\/","title":{"rendered":"Bayesian Inference of Grammars"},"content":{"rendered":"<div class=\"asset-content\">\n<p>Mark Johnson  (Joint work with Sharon Goldwater and Tom Griffiths)<br \/>\nEven though Maximum Likelihood Estimation (MLE) of Probabilistic<br \/>\nContext-Free Grammars (PCFGs) is well-understood (the Inside-Outside<br \/>\nalgorithm can do this efficiently from the terminal strings alone) the<br \/>\ninferred grammars are usually linguistically inaccurate. In order to<br \/>\nbetter understand why maximum likelihood finds poor grammars, this<br \/>\ntalk examines two simple natural language induction problems:<br \/>\nmorphological segmentation and word segmentation. We identify several<br \/>\nproblems with the MLE PCFG models of these problems and propose<br \/>\nHierarchical Dirichlet Process (HDP) models to overcome them. In<br \/>\norder to test these HDP models we develop MCMC algorithms for Bayesian<br \/>\ninference of these models from strings alone. Finally, we discuss to<br \/>\nwhat extent the lessons learnt from these examples can be put into a<br \/>\nunified framework and applied to the general problem of grammar<br \/>\ninduction.<\/p>\n<\/div>\n<p><!-- .asset-content --><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Mark Johnson (Joint work with Sharon Goldwater and Tom Griffiths) Even though Maximum Likelihood Estimation (MLE) of Probabilistic Context-Free Grammars (PCFGs) is well-understood (the Inside-Outside algorithm can do this efficiently from the terminal strings alone) the inferred grammars are usually linguistically inaccurate. In order to better understand why maximum likelihood finds poor grammars, this talk [&hellip;]<\/p>\n","protected":false},"featured_media":194900,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"_classifai_error":"","msr_hide_image_in_river":0,"footnotes":""},"research-area":[],"msr-video-type":[],"msr-locale":[268875],"msr-post-option":[],"msr-session-type":[],"msr-impact-theme":[],"msr-pillar":[],"msr-episode":[],"msr-research-theme":[],"class_list":["post-183106","msr-video","type-msr-video","status-publish","has-post-thumbnail","hentry","msr-locale-en_us"],"msr_download_urls":"","msr_external_url":"https:\/\/youtu.be\/-2XaUwlvZ70","msr_secondary_video_url":"","msr_video_file":"","_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-video\/183106","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-video"}],"about":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-video"}],"version-history":[{"count":0,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-video\/183106\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media\/194900"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=183106"}],"wp:term":[{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=183106"},{"taxonomy":"msr-video-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-video-type?post=183106"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=183106"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=183106"},{"taxonomy":"msr-session-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-session-type?post=183106"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=183106"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=183106"},{"taxonomy":"msr-episode","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-episode?post=183106"},{"taxonomy":"msr-research-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-theme?post=183106"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}