{"id":340575,"date":"2017-05-10T23:34:53","date_gmt":"2017-05-10T15:34:53","guid":{"rendered":"http:\/\/techcrunch.com\/?p=1487545"},"modified":"2017-05-10T23:34:53","modified_gmt":"2017-05-10T15:34:53","slug":"microsoft-maluuba-is-teaching-machines-to-ask-questions","status":"publish","type":"post","link":"https:\/\/people.utm.my\/asmawisham\/microsoft-maluuba-is-teaching-machines-to-ask-questions\/","title":{"rendered":"Microsoft Maluuba\u00a0is teaching machines to ask questions"},"content":{"rendered":"

<\/p>\n

It\u2019s no secret that today\u2019s conversational assistants\u00a0leave a lot to be desired. We spend a lot of time talking about how bad they are at hearing\u00a0what we say and even worse at interpreting it. But one thing that doesn\u2019t get nearly enough attention is the one-directional nature of conversations we have with Siri, Cortana and the rest. These systems don\u2019t know how to ask questions \u2014 real inquiries that reflect a nuanced understanding of what information is known and what information is\u00a0desired.<\/p>\n

Microsoft Maluuba, the research team of PhD\u2019s that Microsoft acquired back in January<\/a> has been hard at work exploring the nexus of machine learning and question generation with the aim of delivering an intelligent personal assistant that surpasses the limited capabilities of today\u2019s market leaders. In a recent paper,<\/a> the team outlined its approach of using recurrent networks to generate\u00a0questions from a body of text.<\/p>\n

The team expanded\u00a0upon existing\u00a0question generators by\u00a0combining supervised learning with reinforcement learning. Reinforcement learning makes use of the economic concept of utility to assign \u201cpoints\u201d for performing what are deemed to be optimal actions. The model is built to maximize this value and over many iterations it\u00a0improves by continuously adjusting its strategy.<\/p>\n