Community Information Service Omid Fatemieh CS 598 CXZ
- Slides: 8
Community Information Service Omid Fatemieh CS 598 CXZ Department of Computer Science University of Illinois at Urbana-Champaign
Current Limitations • Information – Creation: Requires a prof/student to summarize information and discussions and • Update web sites • Write technical reports • Write papers – Storage: Not very well organized and integrated; Scattered in: • Web sites • Personal computers of professors and students • Spaces on department machines
Current Limitations (cont’d) • Information: – Transformation: • Requires a prof/student to gather different sources of information and put them together to have the platform for adding new ideas. – Transportation: • Something should be published or be neatly put on the web site so that it can be somehow transported to other places. • It can also be done through email: – Takes a lot of time – May not be comprehensive
Current Limitations (cont’d) • Information: – Consumption: • Usually requires doing a lot of keyword-driven searchs on different – Professor / students’ web sites – Research groups’ web sites – Related conferences and talks Q: How can we better assist people in academia to: – Speed up the information cycle – Make the above life cycle easier to maintain and use?
Solution Integrated Entity Based Information System with Support for Summarization and Q/A Queries (EBISQA) • Entities can be students, professors, research groups, classes, projects, and even the department itself. • Assigned to each entity: – A knowledge base (profile) which basically keeps track of and summarizes whatever happens to that entity. – Support for Q/A queries on that Entity for people at different levels: • • • The entity itself! People/Entities inside specific groups People inside the department People inside the university People outside the university
Example • Entity: Omid • Assumption: There are speech recognition facilities everywhere backed up by strong NLP techniques and summarization algorithms! • Kind if information that goes into his profile: – Summary of the papers/topics discussed in the classes and seminars he attends – Summary of his meetings with his supervisor – Summary of non-personal messages (e-mails) he sends and receives – All the papers he accesses and projects he works on – Almost everything he does, except for those he flags them as personal.
Solution (cont’d) • Privacy Issues: Support for Q/A queries on his profile at different levels: – Some of them determined by him. – Some of them determined by department policies! • Motivations Problem: • Why I would like to be somehow under “Investigation” all the time? – Will help me in getting more out of my classes, meetings, seminars, by performing searches on my own profile later. – Would help me find a job, if I provide access to my profile to employers. • I would spend $10 M on: – Summarization techniques – NLP algorithms – Q/A queries
Sorry if it was too ambitious! Thank you