The Reap Project: Information Retrieval that Enables Reader-Specific

Vocabulary Instruction

Associate Professor Jamie Callan

Language Technologies Institute (LTI), School of Computer Science, Carnegie Mellon University

Date and time: 11.30am-12.30pm, Friday 21st July, 2006

Venue: 10.08.03 (Building 10, Level 8, Room 3)

Chair: Professor Justin Zobel

Abstract:

The Reap Project is a collaboration between cognitive psychologists and computer scientists that seeks to better understand how children and adults learn new vocabulary.  The project is a combination of software tools that support basic cognitive science research, and an intelligent tutor that provides reading practice.  This talk focuses primarily on the intelligent tutor, which has been used for a year in the English Language Instruction laboratory at the University of Pittsburgh.  It will also briefly discuss other experiments.

The Reap intelligent tutor is based on a mix of information retrieval and statistical natural language technologies.  Each student's vocabulary is modeled using a statistical language model, as is the reference or target vocabulary.  Differences between the two models indicate where a student needs help.  Texts gathered from the Web are used as examples of how target vocabulary is used in the real world ("authentic texts").  Much of the computer science research addresses automatic selection of texts that are educationally appropriate (search, text classification, graphical models, etc), and automatic assessment of student knowledge (e.g., using WordNet, or measures of text similarity).

See http://reap.cs.cmu.edu/ for more information and an animated demo.

About the speaker:

Jamie Callan is an Associate Professor at Carnegie Mellon University's Language Technologies Institute (LTI).  His research covers a variety of topics in Information Retrieval.  His recent research addresses distributed information retrieval / federated search, high speed adaptive information filtering, algorithms that learn information needs by observing user actions, novelty detection, near-duplicate detection, and question answering. His earlier IR research studied architectures for large-scale information retrieval and filtering systems, first generation Web-search systems, integration of text search with relational database systems, and information literacy in K-12 education.

Seminar Organisation

Seminars are free and open to the general public. No booking is necessary. If you are interested in giving a presentation in this seminar series, or to make suggestions for speakers, please contact Xiaodong Li, the seminar co-ordinator.