RSSM: Robust Search over Social Media

Overview

Our goal is to develop effective and robust methods to search and summarize information from social media sites such as Community Question Answering (CQA) portals, online forums, or Wikis. These sites have emerged as viable alternatives to web search for complex or subjective information needs. In particular, CQA sites such as Yahoo! Answers, Baidu Knows, and Naver boast millions of active users that have posted and answered millions of questions. Unfortunately, the quality, completeness, and reliability of the content in such sources vary widely.

This project begins to address fundamental challenges for search and summarization in this setting:

• predictive models of interactions and content creation in social media
• algorithms for searching and organizing of information in social media
• robust evaluation methodology for search over social media