“Making Watson Smarter: From Deep QA to Dynamic QA” (video)

IBM sent Watson to RPI for some education. Part of this education includes the attempt to graduate Watson from mere Deep QA (which sufficed for Jeopardy!) to what we call Dynamic QA. In the former, a computing machine answers questions by merely retrieving answers that are pre-recorded in old data. No one can graduate from college on the strength of providing such answers. In the latter, a computing machine reasons and discovers and problem-solves in real-time, on the spot; this kind of capability earns one a college degree. Some demonstrations involving the robot Baby Watson are provided, and the talk includes elements of prior work devoted to divining what Watson and its successors really are, mathematically speaking.

Click here to view.