Google鈥檚 new Smart Reply artificial intelligence can write e-mails for you
Smart Reply, a feature rolling out this week for Google's Inbox email app, can automatically read incoming emails and suggest appropriate responses. Smart Reply uses machine learning to judge the gist of an email and compose several different reply options.
The Smart Reply feature for Inbox, which will roll out this week, can automatically compose short responses to incoming emails.
Google鈥檚 already investing heavily in machine learning, a type of artificial intelligence in which computers, given a large enough mass of data to chew through, can detect patterns on their own and make judgements about how to handle the information. Machine learning powers Gmail鈥檚 spam filters and the Google Now service, which tries to predict what information a user will find helpful or interesting.
On Tuesday, Google announced that machine learning will be put to use in a new way: to e-mails for you.聽
The 鈥淪mart Reply鈥� feature will roll out later this week to Inbox, Google鈥檚 slightly-experimental e-mail service focused on organization and time management. Smart Reply will automatically parse the text of received e-mails and suggest three responses that can be inserted with a click or a tap.
鈥淔or those emails that only need a quick response, it can take care of the thinking and save precious time spent typing,鈥� software engineer B谩lint Mikl贸 . 鈥淎nd for those emails that require a bit more thought, it gives you a jump start so you can respond right away.鈥�
There鈥檚 a lot going on under the hood of Smart Reply. At the core of the feature is a pair of deep neural networks, which are sets of algorithms that processes data at several layers at once to learn sentence structure, writing style, and tone.
鈥淭hese systems generalize better, and handle completely new inputs more gracefully than brittle, rule-based systems ever could,鈥� Google senior research scientist Greg Corrado .
One neural network parses the text of the incoming e-mail to capture a list of 鈥渢hought vectors鈥� 鈥� essentially, the gist of what鈥檚 being said or asked. If the e-mail contains a question, the second network will compose a simple reply that provides information to answer the query. If an e-mail asks, for example, 鈥淒o you have vacation plans yet,鈥� Smart Reply might suggest as responses, 鈥淣o plans yet,鈥� 鈥淚 just sent them to you,鈥� and 鈥淚鈥檓 working on them.鈥� The system also judges the intent of the proposed responses so that it doesn鈥檛 generate overlapping options that use different phrasing to say the same thing.
Smart Reply also gets more accurate with use, since the networks learn from larger bodies of text. Google鈥檚 early prototype had the propensity to respond to buttoned-down business e-mails with 鈥淚 love you," since the system had identified that as a common phrase and would fall back on it when it was unsure of how to respond. The team fixed that issue by having Smart Reply check to see how often phrases had been used previously.