• Home

• People

• Labs

• For Students

• Publications

• Software

• Events

• Locations

• Internal

Natural Language Text Processing Lab

The Natural Language Text Processing Lab at Columbia, which was established by Prof. Kathleen R. McKeown, is pursuing a broad range of research, including natural language generation, question answering, concept to speech generation, summarization of news, meetings and technical literature, statistical language modelling and digital libraries.

A number of the applications we are developing address the problems and opportunities that come with the rapid growth of the World Wide Web. Among these are a system to produce multimedia briefings or updates of on-line news articles and a domain-independant system to generate summaries of a number of articles on a related subject.

In a number of projects, we collaborate with the medical and journalism schools at Columbia, with several other groups within the Computer Science Department, and have enjoyed good relations with researchers at IBM, AT&T and Google.

In more basic research, we are exploring how to apply machine learning techniques to the natural language problem, conducting statistical analyses of large text corpora and applying the results to text generation, and trying to develop methods of generating speech from concepts. Several on-going projects are looking at innovative ways to extract information for domain-independent summaries and to improve the readability of the summaries themselves.

Quick Links

To learn about our current projects, visit our Projects Page. Undergraduate and master students interested in research projects should check out our Ads for Student Research Projects in NLP.

People includes the current faculty, research staff, students and alumni of the group. The PhD student pages indicate other research topics not included in the Project section.

Newsblaster, Columbia News Tracking and Summarizing tool, provides automatically generated summaries of the main news articles every day.

In Tools you will find downloadable packages developed by our group.