The Computer Science Department at the University of Rochester has a long history of innovative work in areas related to task-oriented language use, including:

• spoken dialogue systems
• discourse analysis
• planning and plan recognition
• knowledge representation and reasoning

Many of our past and present projects are listed below.

Spoken Dialogue

Our research on spoken dialogue systems centers around a long-term effort to develop an intelligent planning assistant that is conversationally proficient in natural language. The goal is a fully integrated system involving online spoken and typed natural language together with graphical displays and GUI-based interaction. Our current research on spoken dialogue involves the TRIPS system.

The research runs along two tracks. First, the projects support Ph.D. thesis research in natural language processing, speech interpretation, discourse processing, planning, and reasoning under uncertainty. Second, we periodically pull the research threads together and produce limited but working demonstration systems based on the theories. The hope is that this incremental development is converging towards the ultimate goal of a fully conversant system.

• PLOW
• TRIPS
• TRAINS
• Comparison of TRAINS and TRIPS
• Intelligent Interactive Assistants
• ARGOT

Discourse Analysis

Our research in this area includes collecting and annotating task-oriented spoken dialogs (human-human and human-machine), and developing annotation tools for spoken dialogue.

• DAMSL
• Annotating argumentation acts in dialog
• TRAINS Corpus
• Monroe Corpus

Natural Language Understanding

• Geospatial Natural Language Understanding
  • PURSUIT Corpus

Planning and Plan Recognition

In the past, the University of Rochester has been the site of research in temporal reasoning for planning, plan recognition and reasoning with planning. Current research in planning takes place in the context of the TRIPS project, and as collaboration with visiting researchers. See our publications list for details.

• In conjunction with Alfonso Gerevini (U. Brescia), Fast temporal reasoning and planning
• Statistical Goal Recognition
  • Linux Plan Corpus
  • Monroe Plan Corpus

Knowledge representation and reasoning

Past and present research in this area includes research in logics of time and events, research in representation of world knowledge, and research into default logics.

• EPILOG
• Mining text for general world knowledge (Lenhart Schubert, Henry Kyburg, Greg Carlson)
• Plausible inference, classification, and default logics (Henry Kyburg and students)
• The Rhetorical Knowledge Representation System (Rhet) and its successor, Shocker (James Allen and Brad Miller)