This post has been republished via RSS; it originally appeared at: Channel 9.
Project Conversation Learner enables you to build and teach task-oriented conversational interfaces that learn directly from example interactions. Whereas traditional approaches to dialog control rely on a static flowchart or set of rules, Conversation Learner chooses responses using an end-to-end neural network which can learn nuanced dialog sequences. This machine learning core can be paired with code to add business rules, domain knowledge, and common sense. Overall, this approach reduces manual coding of dialogue control logic, and enables business owners or domain experts to contribute to a conversational interface without prior machine learning knowledge. Whether it's deployed as part of a bot, smart device, or intelligent agent, with Conversation Learner, you can rapidly create new conversational skills and competencies, and quickly improve quality. Bots built using this approach can be readily deployed via the Microsoft Bot Framework, or as standalone services.
