Univerity of Delaware Department of Computer Science

Graduate Program in Artificial Intelligence and Natural Language Processing

Artificial Intelligence is a major areas of research and graduate study in the Department of Computer and Information Sciences at the University of Delaware. Current research interests center on natural language generation and understanding, multi-agent systems, neural networks, planning and plan recognition, user modeling, cooperative distributed problem solving, and rehabilitation engineering. A Cognitive Science Program brings in well-known speakers from throughout North America and serves as a forum for joint research efforts with faculty from the departments of Educational Studies, Linguistics, and Psychology. Close interaction with the Center for Applied Science and Engineering in Rehabilitation of the University of Delaware and the AI duPont Institute provides opportunities for research in natural language processing, speech synthesis, and robotics.

Current Faculty

Members of our AI group serve on the editorial boards of Computational Linguistics Journal, International Journal of Man-Machine Studies, Autonomous Agents and Multi-Agent Systems Journal, User Modeling and User-Adapted Interaction Journal, Machine-Mediated Learning, and Visible Language. In addition, they have served on the program committees for major conferences in artificial intelligence, natural language processing, and multi-agent systems.

Graduate Students

Click here for a current list of our graduate students.

Research Interests

Knowledge representation and user modeling are areas of active research interest. Precisely how knowledge is derived from a sequence of observed actions (including a sequence of utterances in dialogue) and how this knowledge can be represented and reasoned with are major unsolved problems. Planning, problem-solving, and plan recognition are important components of intelligent systems, especially systems that must react to their environment or interact with other agents, and several researchers are pursuing work in these areas. In addition, increasing attention is being devoted to neural networks as a computational and cognitive paradigm, and members of the computer science department are concerned with neural network propagation algorithms and approaches to language processing.

Multi-Agent Systems and Distributed Artificial Intelligence research deals with the issues that arise when groups or societies of autonomous agents (usually computer programs but sometimes people too) interact to solve problems. These agents may be self-interested, or cooperating to solve a shared problem. Important issues include reasoning about the knowledge and beliefs of other agents, communication and negotiation, and coordination and control. Furthermore, in many real-world problems agents have limited computational resources available to them, and so must forgo optimal solutions for satisficing solutions.

Natural Language Processing is a major research area within the department and is aided by interaction with faculty in the Department of Linguistics and the Department of Psychology, and by graduate linguistics and psychology courses. Efforts focus on developing theoretical formalisms to describe linguistic phenomena, devising strategies for enabling machines to communicate effectively in a natural language such as English, and developing augmentative communication systems for people with severe speech-impairments.

Course Offerings

CISC 681 Introduction to Artificial Intelligence
CISC 689 Topics: Neural Networks
CISC 809 Advanced Topics: Non-Standard Logics
CISC 881 Knowledge-Based Systems
CISC 882 Natural Language Processing
CISC 883 Natural Language Generation
CISC 884 Knowledge Representation
CISC 885 Discourse and Dialogue
CISC 889 Advanced Topics: Unification-Based Grammatical Formalisms
CISC 889 Advanced Topics: Planning and Reasoning
CISC 889 Advanced Topics: Multi-Agent Systems
CISC 889 Advanced Topics: Internet Information Gathering
EDST 633 Introduction to Computer-based Learning
LING 609 Syntax I
LING 610 Syntax II
LING 691 Semantics
LING 696 Psycholinguistics
PSYC 640 Text Comprehension
PSYC 641 Visual Cognition
PSYC 642 Mental Representation and Memory
PSYC 667 Computer Models of Cognition

Research Projects

The following is a brief description of some of the current AI-related research projects:

Alphabet Superhighway (Venezky): This WWW project assists schools in using technology for instruction by providing exemplary elementary and secondary school student projects, ideas, and other resources. Funded by the U.S. Department of Education and by the Joyce Foundation to bring the project to inner-urban schools in the midwest.

Automated Problem Representation (Chester): Much of the difficulty in solving problems is in putting them into representations that make their solution tractable. This project aims to automate the problem representation process for problems that have been stated in natural language accompanied by diagrams. It combines natural language processing and computer vision to solve problems without any of the preprocessing that is usually done by humans.

Collaborative Consultation Dialogues (Carberry): This project investigates the modeling of collaborative consultation dialogues within a plan-based framework. Strategies are being developed for recognizing problem-solving and communicative intentions and for handling negotiation subdialogues. In addition, a plan-based response generation component is being formulated. Funded by NSF: $179,916 for three years.

Computational Organization Design (Decker): Human organizational theory has many different things to say about what organizations are and why and how different organizational forms come about. This project studies what we can learn from these theories, and how they can apply to the construction of new organizations that rely on active computational components. Furthermore, as we build distributed AI systems with agents that have both limited resources and their own, possibly conflicting goals and intentions, human organizational theories can be shown to provide useful organizing structures and concepts.

Computer Aided Writing Tool for Deaf Individuals (McCoy): The goal of this project is to develop a computer program that will analyze text written by a deaf person, and suggest corrections to that text. The hypothesis that drives the development is that American Sign Language (ASL) is a native language to some deaf people, and thus English should be thought of/taught as a second language for these people. Initial research funded by NSF: $45,371.

Design and Analysis of Coordination Mechanisms (Decker): The Generalized Partial Global Planning approach to multi-agent system coordination allows the use of many different mechanisms to manage interdependencies in human and/or computational agent activities. This project studies the characteristics of existing coordination mechanisms, and develops new ones, for a range of distributed applications such as hospital scheduling, concurrent engineering, distributed sensor interpretation, and information gathering.

DIALS (Chester): Software for natural language parsing, plan recognition, question answering, and text generation are being combined in the Delaware Intelligent Advising Language System, which will serve as a testbed for our theories of natural language processing.

Discourse Initiative in Intelligent Tutoring Systems (Carberry): Current intelligent tutoring systems do not accommodate and take advantage of the many ways in which students can respond during training. This research will investigate the kinds of student initiative that intelligent tutoring systems should encourage, how student initiative can be accommodated within the limits of current technology, how the information provided by student initiative can be utilized by the system, and the extent to which student initiative impacts the effectiveness of training systems.

Effective Information Delivery in a Real-time Decision Support System (Carberry). This project, pursued in conjunction with researchers at the University of Pennsylvania and the Medical College of Pennsylvania, is investigating mechanisms for effective delivery of information by a medical decision support system during real-time patient care. One component of our work has been to develop a message planner that produces integrated messages from individual critiques each of which is designed to achieve its own communicative goal. Future work will address other communication problems. This project is funded by the National Library of Medicine: approximately $200,000 for three years.

Gesture, Intonation, and Discourse Acts (Carberry, Chester, and Kambhamettu): Prosody and gesture are mechanisms that speakers use to convey their intentions but that have not been adequately taken into account by current models of discourse. This project will develop algorithms for discourse act recognition that utilize evidence from intonation and gesture. In addition, the project will investigate how statistical methods should be incorporated into a recognition algorithm.

Grammar Formalisms (Vijayashanker): We have been investigating Tree-Adjoining Grammars, a widely used grammatical system. In recent work, we have also been investigating some related formalisms, in particular that of D-Tree Grammars, and evaluating these formalisms especially from the point of view of descriptive adequacy, and parsing efficiency.

Intelligent Training Systems (Carberry): The goal of this project is an intelligent training system in the domain of medical trauma care. As a first step, we have developed a system, TraumaCASE, for automatically generating medical cases of the appropriate level of difficulty by reasoning with the declarative knowledge base of an existing medical decision support system. In the future, we will extend TraumaCASE to provide incremental tailored case generation and will develop an intelligent tutoring sytem for user-adapted training in the management of trauma cases.

Multi-Agent Information Gathering (Decker): A large amount of useful information is currently available on the Internet, but finding that information and filtering it can be a daunting task both because of the amount of data and he fact that it is constantly changing. This project seeks to bring multi-agent system approaches to bear on this problem in application areas such as computer product awareness and financial portfolio management.

Parsing (Vijayashanker): We are investigating development of parsing algorithms for a range of grammar formalisms and are considering algorithmic techniques that have not been widely studied in this context. We are also studying the use of underspecification during the parsing process in an attempt to handle the ambiguity problem in a more efficient manner.

Rehabilitation Engineering Research Center on Augmentative Communication (Bunnell, McCoy): This center includes a number of projects pursued by faculty and students in the computer science department, including

Funded, along with other research on augmentative communication, by the National Institute on Disability and Rehabilitation Research of the U.S. Dept. of Education: $700,000 per year for five years.

Rehabilitation Engineering Research Center in Robotics to Enhance the Functioning of Individuals with Disabilities (Foulds, Harwin, Rahman, Chester): This center includes a number of projects pursued by faculty and students in the computer science department, including

Funded, along with other research on robotics, by the National Institute on Disability and Rehabilitation Research of the U.S. Dept. of Education; $700,000 per year for five years.

Sentence Generation Architecture (McCoy, Vijayashanker): This project, on sentence generation, involves bringing together two different representations of linguistic knowledge, viz., a functional perspective and a syntactic perspective. We have developed an architecture that combines these perspectives in a way that reaps the benefits that each brings to the generation process. Current work on the project involves issues in the representation of the lexicon and syntactic structures, and multi-lingual generation.

Towards a Logical Formulation of Tree-based Formalisms (Vijayashanker): This project involves developing a logical calculus in which various notions of trees and tree-rewriting can be stated and reasoned about. We are investigating the choice of appropriate primitives that will be suited towards succinctly describing linguistic primitives. We are also investigating the use of this calculus in providing a workbench to develop and organize a large-scale grammar in a space-efficient manner.

Use of Focus in Natural Language Generation (McCoy): In order for a computer system to generate written text, it must have some method for deciding what to say next and for deciding when it is appropriate to use a pronoun. The way that focus shifts through discourse, and how focus can affect pronoun resolution are part of the information captured in a focusing framework. In this work we look at developing appropriate focusing frameworks and identifying how focus shifts in a coherent discourse.

Sample Recent Publications

Keith Decker, Anandeep Pannu, Katia Sycara, and Mike Williamson. Designing Behaviors for Information Agents. In Proceedings of the First International Conference on Autonomous Agents, Marina del Rey, February 1997.

Kathleen F. McCoy and Patrick W. Demasco. Some Interface Issues in Developing Intelligent Communication Aids for People with Disabilities. In Proceedings of the 1997 International Conference on Intelligent User Interfaces, IUI97, Orlando, Florida, January, 1997.

Carberry, S. and J. Chu-Carroll and L. Lambert. Modeling Intention: Issues for Spoken Language Dialogue Systems. Proceedings of the 1996 International Symposium on Spoken Dialogue, pp. 13-24, 1996.

S. Chen, Z. Kazi, M. Beitler, M. Salganicoff, D. Chester, and R. Foulds. Gesture-Speech Based HMI for a Rehabilitation Robot. Proceedings of IEEE Southeastcon'96, Tampa, Florida, pp. 29-36, April 1996.

Keith Decker, Katia Sycara, and Mike Williamson. Modeling Information Agents: Advertisements, Organizational Roles, and Dynamic Behavior. Working Notes of the AAAI-96 workshop on "Agent Modeling". AAAI Report WS-96-02. August 1996.

Z. Kazi, M. Beitler, M. Salganicoff, S. Chen, D. Chester, and R. Foulds. Multimodally Controlled Intelligent Assistive Robot. Proceedings of RESNA 1996, Salt Lake City, Utah, pp. 348-350, June 1996.

Kathleen F. McCoy, Christopher Pennington, Linda Z. Suri. English Error Correction: A Syntactic User Model Based on Principled `Mal-Rule' Scoring. In Proceedings of the Fifth International Conference on User Modeling, Kailua-Kona, Hawaii, January, 1996.

Kathleen F. McCoy, Christopher A. Pennington, and Linda Z. Suri. Considering the Effects of Second Language Learning on Generation. In Proceedings of the 8th International Natural Language Generation Workshop, Herstmonceux Castle, Sussex, UK, June, 1996.

Katia Sycara, Keith Decker, Anandeep Pannu, Mike Williamson, and Dajun Zeng. Distributed Intelligent Agents. IEEE Expert, 11(6):36--46, December 1996.

K. Vijay-Shanker. The Use of Domination Statements in Tree-Adjoining Grammars and Related Formalisms, Proceedings of 12th International Conference on Natural Languages and Formal Languages, La Seu D'Urgell, Spain, September 1996.

R. Backofen, J. Rogers, and K. Vijay-Shanker. A First-order Axiomatization of the Theory of Finite Trees, Journal of Logic, Language, and Information, June 1995, 4:5-39.

Chu-Carroll, J. and S. Carberry. Generating Information-Sharing Subdialogues in Expert-User Consultation. Proceedings of the 14th International Joint Conference on Artificial Intelligence, pp. 1243-1250, Aug 1995.

Chu- Carroll, J. and S. Carberry. Response Generation in Collaborative Negotiation. Proceedings of the 33rd Annual Meeting of the Association for Computational Linguistics, pp. 136-143, June 1995.

Chu-Carroll, J. and S. Carberry. Communication for Conflict Resolution in Multi-Agent Collaborative Planning. Proceedings of the First International Conference on Multi-Agent Systems, pp. 49-56, July 1995.

Keith S. Decker and Victor R. Lesser. Designing a Family of Coordination Algorithms. In Proceedings of the First International Conference on Multi-Agent Systems, San Francisco, July 1995.

R. Kasper, K. Netter, and K. Vijay-Shanker. Compilation of HPSG to TAG. Proceedings of 33rdAnnual Conference of the Association for Computational Linguistics, Cambridge, MA, July 1995.

O. Rambow, K. Vijay-Shanker, and D. J. Weir. A Parsing Algorithm for D-Tree Grammars, Proceedings of 5th International Workshop on Parsing Technologies, Prague, August 1995.

Keith S. Decker. Distributed Artificial Intelligence Testbeds. In G. O'Hare and N. Jennings, editors, Foundations of Distributed Artificial Intelligence. Chapter 3. Wiley Inter-Science, 1995.

Chu-Carroll, J. and S. Carberry. A Plan-Based Model for Response Generation in Collaborative Task-Oriented Dialogues. Proceedings of the Twelfth National Conference on Artificial Intelligence, pp. 799-805, 1994.

Kathleen F. McCoy, Patrick W. Demasco, Mark A. Jones, Christopher A. Pennington, Peter B. Vanderheyden, Wendy M. Zickus. A Communication Tool for People with Disabilities: Lexical Semantics for Filling in the Pieces. In Proceedings of The First Annual International ACM/SIGCAPH Conference on Assistive Technologies, Los Angeles, CA, 1994.

K. Vijay-Shanker and D. J. Weir. Equivalence of Some Extensions of Context-free Grammars, Mathematical Systems Theory, Volume 27, December 1994.

Linda Z. Suri and Kathleen F. McCoy RAFT/RAPR and Centering: A Comparison and Discussion of Problems Related to Processing Complex Sentences. Computational Linguistics, Vol. 20, No. 2, Squibs and Discussions, pp. 301-317, June, 1994.

Keith S. Decker and Victor R. Lesser. An approach to analyzing the need for meta-level communication. In Proceedings of the Thirteenth International Joint Conference on Artificial Intelligence, Chambéry, August 1993.

Green, N. and S. Carberry. A Hybrid Reasoning Model for Indirect Answers. Proceedings of the 32nd Annual Meeting of the Association for Computational Linguistics, pp. 58-65, 1994.

Sandra Carberry and Alan Pope. Plan Recognition Strategies for Language Understanding. International Journal of Man-Machine Studies, November, 1993.

K. Vijay-Shanker and David Weir. Parsing Some Constrained Grammar Formalisms. Computational Linguistics, Vol 19, Number 4, 1993.

Linda Suri and Kathleen F. McCoy. Correcting Discourse-Level Errors in a CALL System for Second Language Learners. Computer Assisted Language Learning, 6(3), pp. 215-231, 1993.

Return to CIS Home Page

This page last edited March 29, 1998.