What if you could ask your email client, “Who sent me the link with the latest financial report?” Automated question answering is a human-machine interaction to extract information from data using natural language. This general capability can take many forms, but one of the most exciting developments has been question answering from unstructured text data, including the massive amounts of information contained in emails, social media posts, blogs, log files, financial statements -- and the list goes on. Thanks to a series of advances in deep learning techniques in the past two years, question answering capabilities have grown rapidly, and while still emerging, it’s the perfect time to examine how this technology works, when it works well, and where it might still fall short.
In this webinar we’ll cover
General architecture of modern QA systems
Deep learning techniques for QA
Guidance on applying these techniques to a practical use case
We’ll also do a live demonstration of a QA prototype that we built. You won’t want to miss it, so we hope to see you there!