Michael Sol Silver
MIT EECS | Quick Undergraduate Research and Innovation Scholar
Question Answering from Wikipedia
- Computer Systems
Question answering is a growing field of interest as personal assistants become more commonplace. Wikipedia has become an ever-growing source of rich information for question answering. The InfoLab developed a question answering system called START and created a dedicated system called WikipediaBase to extract information from Wikipedia. Our research will aim at expanding the current abilities of WikipediaBase and enable START to continually learn more information as Wikipedia grows. Our system and the models of information extraction are general strategies that serve as inspiration for commercial systems such as Siri, Cortana, and even some of Google’s question answering functionality. Research on further advancing START can lead to improvements in these and many other systems.
“I am participating in SuperUROP to continue my journey in developing systems that support AI. I’ve developed a love for systems after taking Courses 6.033 (Computer Systems Engineering) and 6.858 (Computer Systems Security), and I have been involved with AI research for my entire time at MIT. Combining the two is incredibly exciting.”