{"id":32376,"date":"2018-10-29T13:47:41","date_gmt":"2018-10-29T08:17:41","guid":{"rendered":"https:\/\/blog.forumias.com\/?p=32376"},"modified":"2018-10-29T13:47:41","modified_gmt":"2018-10-29T08:17:41","slug":"googles-new-ai-system-can-atriculate-like-humans","status":"publish","type":"post","link":"https:\/\/forumias.com\/blog\/googles-new-ai-system-can-atriculate-like-humans\/","title":{"rendered":"Googles new AI system can atriculate like humans\u00a0"},"content":{"rendered":"<p><a href=\"http:\/\/www.thehindu.com\/sci-tech\/technology\/googles-new-ai-system-can-articulate-like-humans\/article22347293.ece\"><strong>Googles new AI system can atriculate like humans<\/strong><\/a><strong>\u00a0<\/strong><\/p>\n<p><strong>Context <\/strong><\/p>\n<p>In a major step towards its \u201cAI first\u201d dream, Google has developed a text-to-speech artificial intelligence (AI) system that will confuse you with its human-like articulation<\/p>\n<p><strong>Tacotron 2<\/strong><\/p>\n<p>The tech giant\u2019s text-to-speech system called \u201cTacotron 2\u201d delivers an AI-generated computer speech that almost matches with the voice of humans, technology<\/p>\n<p><strong>How the system works?<\/strong><\/p>\n<ul>\n<li>The system first creates a spectrogram of the text, a visual representation of how the speech should sound<\/li>\n<li>That image is put through Google\u2019s WaveNet algorithm, which uses the image and brings AI closer than ever to mimicking human speech. It can easily learn different voices and even generates artificial breaths<\/li>\n<\/ul>\n<p><strong>Mean Opinion Score (MOS)<\/strong><\/p>\n<p>\u201cOur model achieves a mean opinion score (MOS) of 4.53 comparable to a MOS of 4.58 for professionally recorded speech,\u201d the researchers were quoted as saying<\/p>\n<p><strong>What is MOS?<\/strong><\/p>\n<p>It\u00a0is a numerical method of expressing voice and video quality<\/p>\n<ul>\n<li>MOS gives a numerical indication of the perceived quality of the media received after being transmitted and eventually compressed\u00a0using codecs<\/li>\n<li>MOS is expressed in one number, from 1 to 5, 1 being the worst and 5 the best. MOS is quite subjective, as it is based figures that result from what is perceived by people during tests. However, there are\u00a0software\u00a0applications that measure MOS on networks<\/li>\n<\/ul>\n<p>MOS values<\/p>\n<p><strong>The Mean Opinion Score Values<\/strong><\/p>\n<p>Taken in whole numbers, the numbers are quite easy to grade.<\/p>\n<p><strong>5<\/strong>\u00a0&#8211; Perfect. Like face-to-face conversation or radio reception<\/p>\n<p><strong>4<\/strong>\u00a0&#8211; Fair. Imperfections can be perceived, but sound still clear. This is (supposedly) the range for cell phones.<\/p>\n<p><strong>3<\/strong>\u00a0&#8211; Annoying<\/p>\n<p><strong>2<\/strong>\u00a0&#8211; Very annoying. Nearly impossible to communicate.<\/p>\n<p><strong>1<\/strong>\u00a0&#8211; Impossible to communicate<\/p>\n<p><strong>AI first<\/strong><\/p>\n<p>At Google I\/O 2017 developers conference, the company\u2019s CEO announced that the internet giant was shifting its focus from mobile-first to \u201cAI first\u201d and launched several products and features, including Google Lens, Smart Reply for Gmail and Google Assistant for iPhone<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Googles new AI system can atriculate like humans\u00a0 Context In a major step towards its \u201cAI first\u201d dream, Google has developed a text-to-speech artificial intelligence (AI) system that will confuse you with its human-like articulation Tacotron 2 The tech giant\u2019s text-to-speech system called \u201cTacotron 2\u201d delivers an AI-generated computer speech that almost matches with the&hellip; <a class=\"more-link\" href=\"https:\/\/forumias.com\/blog\/googles-new-ai-system-can-atriculate-like-humans\/\">Continue reading <span class=\"screen-reader-text\">Googles new AI system can atriculate like humans\u00a0<\/span><\/a><\/p>\n","protected":false},"author":61,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"jetpack_post_was_ever_published":false,"footnotes":""},"categories":[555],"tags":[],"class_list":["post-32376","post","type-post","status-publish","format-standard","hentry","category-test-1","entry"],"jetpack_featured_media_url":"","views":{"total":0,"cached_at":"","cached_date":1704878714},"jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/forumias.com\/blog\/wp-json\/wp\/v2\/posts\/32376","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/forumias.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/forumias.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/forumias.com\/blog\/wp-json\/wp\/v2\/users\/61"}],"replies":[{"embeddable":true,"href":"https:\/\/forumias.com\/blog\/wp-json\/wp\/v2\/comments?post=32376"}],"version-history":[{"count":0,"href":"https:\/\/forumias.com\/blog\/wp-json\/wp\/v2\/posts\/32376\/revisions"}],"wp:attachment":[{"href":"https:\/\/forumias.com\/blog\/wp-json\/wp\/v2\/media?parent=32376"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/forumias.com\/blog\/wp-json\/wp\/v2\/categories?post=32376"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/forumias.com\/blog\/wp-json\/wp\/v2\/tags?post=32376"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}