{"id":16153,"date":"2017-05-24T23:44:59","date_gmt":"2017-05-24T23:44:59","guid":{"rendered":"http:\/\/www.kurzweilai.net\/?p=301313"},"modified":"2017-05-26T18:54:13","modified_gmt":"2017-05-26T18:54:13","slug":"how-googles-smart-reply-is-getting-smarter","status":"publish","type":"post","link":"https:\/\/hoo.central12.com\/fugic\/2017\/05\/24\/how-googles-smart-reply-is-getting-smarter\/","title":{"rendered":"How Google’s ‘smart reply’ is getting smarter"},"content":{"rendered":"
\"\"

(credit: Google Research)<\/p><\/div>\n

Last week, KurzweilAI<\/em> reported<\/a> that Google is rolling out an enhanced version of its \u201csmart reply<\/a>\u201d machine-learning email software to \u201cover 1 billion Android and iOS users of Gmail\u201d — quoting Google CEO Sundar Pichai.<\/p>\n

We noted that the new smart-reply version is now able to handle challenging sentences like \u201cThat interesting person at the cafe we like gave me a glance,\u201d as Google research scientist Brian Strope and engineering director Ray Kurzweil noted in a Google Research blog post<\/a>.<\/p>\n

But \u201cgiven enough examples of language, a machine learning approach can discover many of these subtle distinctions,\u201d they wrote.<\/p>\n

How does it work? “The content of language is deeply hierarchical, reflected in the structure of language itself, going from letters to words to phrases to sentences to paragraphs to sections to chapters to books to authors to libraries, etc.,” they explained.<\/p>\n

So a hierarchical approach to learning “is well suited to the hierarchical nature of language. We have found that this approach works well for suggesting possible responses to emails. We use a hierarchy of modules, each of which considers features that correspond to sequences at different temporal scales, similar to how we understand speech and language.”*<\/p>\n

Simplfying communication<\/strong><\/p>\n

\u201cWith Smart Reply, Google is assuming\u00a0users want to offload the burdensome task of communicating with one another to our more efficient counterparts,\u201d\u00a0says<\/a> Wired<\/em> writer Liz Stinson.<\/p>\n

\u201cIt\u2019s not wrong. The company says the machine-generated replies already account for 12 percent of emails sent; expect that number to boom once everyone with the Gmail app can send one-tap responses.<\/p>\n

\u201cIn the short term, that might mean more stilted conversations in your inbox. In the long term, the growing number of people who use these canned responses is only going to benefit Google, whose AI grows smarter with every email sent.\u201d<\/p>\n

Another challenge is that our emails, particularly from mobile devices, \u201ctend to be riddled with idioms [such as urban lingo] that make no actual sense,\u201d suggests<\/a> Washington Post<\/em> writer Hayley Tsukayama. \u201cThings change depending on context: Something ‘wicked’ could be good or very bad, for example. Not to mention, sarcasm is a thing.<\/p>\n

\u201cWhich is all to warn you that you may still get a wildly random and even potentially inappropriate suggestion — I once got an ‘Oh no!’ suggestion to a friend\u2019s self-deprecating pregnancy announcement, for example. If the email only calls for a one- or two-sentence response, you\u2019ll probably find Smart Reply useful. If it requires any nuance, though, it\u2019s still best to use your own human judgment.\u201d<\/p>\n

* The initial release of Smart Reply encoded input emails word-by-word with a long-short-term-memory<\/a> (LSTM) recurrent neural network, and then decoded potential replies with yet another word-level LSTM. While this type of modeling is very effective in many contexts, even with Google infrastructure, it\u2019s an approach that requires substantial computation resources. Instead of working word-by-word, we found an effective and highly efficient path by processing the problem more all-at-once, by comparing a simple hierarchy of vector representations of multiple features corresponding to longer time spans. — Brian Strope and Ray Kurzweil, <\/em>Google Research Blog.<\/em><\/p>\n","protected":false},"excerpt":{"rendered":"

Last week, KurzweilAI reported that Google is rolling out an enhanced version of its “smart reply” machine-learning email software to “over 1 billion Android and iOS users of Gmail” — quoting Google CEO Sundar Pichai. We noted that the new smart-reply version is now able to handle challenging sentences like “That interesting person at the […]<\/p>\n","protected":false},"author":13,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[46,43],"tags":[],"class_list":["post-16153","post","type-post","status-publish","format-standard","hentry","category-airobotics","category-news"],"_links":{"self":[{"href":"https:\/\/hoo.central12.com\/fugic\/wp-json\/wp\/v2\/posts\/16153"}],"collection":[{"href":"https:\/\/hoo.central12.com\/fugic\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/hoo.central12.com\/fugic\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/hoo.central12.com\/fugic\/wp-json\/wp\/v2\/users\/13"}],"replies":[{"embeddable":true,"href":"https:\/\/hoo.central12.com\/fugic\/wp-json\/wp\/v2\/comments?post=16153"}],"version-history":[{"count":4,"href":"https:\/\/hoo.central12.com\/fugic\/wp-json\/wp\/v2\/posts\/16153\/revisions"}],"predecessor-version":[{"id":16219,"href":"https:\/\/hoo.central12.com\/fugic\/wp-json\/wp\/v2\/posts\/16153\/revisions\/16219"}],"wp:attachment":[{"href":"https:\/\/hoo.central12.com\/fugic\/wp-json\/wp\/v2\/media?parent=16153"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/hoo.central12.com\/fugic\/wp-json\/wp\/v2\/categories?post=16153"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/hoo.central12.com\/fugic\/wp-json\/wp\/v2\/tags?post=16153"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}