{"id":569514,"date":"2019-02-24T10:44:56","date_gmt":"2019-02-24T18:44:56","guid":{"rendered":"https:\/\/www.microsoft.com\/en-us\/research\/?post_type=msr-project&p=569514"},"modified":"2019-06-17T13:49:02","modified_gmt":"2019-06-17T20:49:02","slug":"multi-domain-task-completion-dialog-challenge","status":"publish","type":"msr-project","link":"https:\/\/www.microsoft.com\/en-us\/research\/project\/multi-domain-task-completion-dialog-challenge\/","title":{"rendered":"Multi-Domain Task-Completion Dialog Challenge"},"content":{"rendered":"
As part of the Eighth Dialog System Technology Challenge (DSTC8<\/a>), Microsoft Research and Tsinghua University are hosting a track intended to foster progress in two<\/strong> important aspects of dialog systems: dialog complexity and scaling to new domains. For this DSTC8 track, there are two<\/strong> tasks you can compete in (see below). The challenge runs from\u00a0June 17, 2019 – October 6, 2019.<\/strong><\/p>\n <\/p>\n Task 1<\/strong> –\u00a0 There is increasing interest in building complex bots that span over multiple sub-domains to accomplish a complex user goal such as travel planning. Travel planning may include sub-domains like hotels, restaurants, tourist attractions, and so on. To advance state-of-the-art technologies for handling complex dialogs, we offer a timely task focusing on multi-domain end-to-end task completion dialog.<\/p>\n Sign up to participate in Task 1 at https:\/\/aka.ms\/dstc8-task1<\/a>.<\/p>\n <\/p>\n Task 2<\/strong> – Neural dialog systems require very large datasets to learn how to output consistent and grammatically-correct sentences. This need for large datasets makes it extremely hard to scale out the system to new domains with limited in-domain data. With Task 2, our goal is to investigate whether sample complexity can decrease with time. In other words, the goal of Task 2 is to investigate whether a dialog system that was trained on a large corpus can learn to converse about a new domain given a much smaller in-domain corpus.<\/p>\n Sign up to participate in Task 2 at https:\/\/aka.ms\/dstc8-task2<\/a>.<\/p>\n