# This file contains your research projects. # # Each entry in this file is formatted as follows: # # %TI Title of project # %AU Your name, other faculty name, other faculty name # %AV URL for this project # %AB Brief abstract for this project (100 words). # # Each field should be on one line, except %AB which can run over # several lines. You may use mailto:YOUR-EMAIL-ADDRESS for the URL # in the %AV field if the project does not have a Web page. %TI Abductive Explanation %AU Ashwin Ram %AV http://www.cc.gatech.edu/faculty/ashwin/projects/explanation.html %AB Most comprehension tasks, such as story understanding, require that the reasoner be able to construct explanations of observed phenomena. In order to explain a given situation, the reasoner must be able to identify what needs to be explained, to use its past experiences to construct hypotheses about the current situation, and to verify its hypotheses. Furthermore, the reasoner must be able to evaluate the utility of proposed explanations with respect to its goals or tasks. A case-based reasoning approach allows a reasoner to perform these tasks and to improve its own ability to explain through experience. Implemented systems include AQUA, a story understanding system that asks questions and attempts to answer them in its quest to explain and understand unusual newspaper stories. %TI Case-Based Design Aids %AU Janet Kolodner %AV http://www.cc.gatech.edu/aimosaic/faculty/kolodner/cbda.html %AB Projects include ARCHIE, a case-based design aid for the conceptual design of buildings; MIDAS, memory for the initial design of aircraft subsystems; and Design-MUSE, a shell that eases construction of case-based design aids. %TI Cognitive Multimedia %AU Richard Catrambone, Mark Guzdial, Ashwin Ram, Preetha Ram, Mimi Recker, John Stasko %AV http://www.cc.gatech.edu/faculty/ashwin/projects/cognitive-multimedia/ %AB The Cognitive Multimedia Project is an on-going research effort to develop multimedia systems to support human learning and problem solving, based on cognitive principles and guidelines from cognitive science. Recent advances in computing technology have the potential to make truly multimedia educational and support environments possible. Such environments provide individualized and interactive support for problem-solving and learning activities of a human user, and have the potential to enhance education for students, training for novices, or on-the-job support for experts. However, in order to fully exploit the promise of multimedia and its related technological capabilities, we need a cognitive theory of human learning and problem solving and an associated set of design principles for the development of multimedia support environments. A system based on these principles would support not only access to multiple types of information but also effective reasoning and learning activities. Implemented systems include AlgoNet, a multimedia system that supports hypermedia information access and constructive activities for self-paced learning in computer and engineering disciplines; ChemLab, a multimedia system that supports constructive learning activities for an undergraduate chemistry course; and WALTS, a multimedia workspace for aiding and training troubleshooters in an engineering domain. %TI Cognitive Science %AV http://www.cc.gatech.edu/cogsci/ %AB We are a multidisciplinary group of faculty, research scientists, and students with several points of interdisciplinary connection in research and in teaching. A distinctive, unifying focus of the program is the study of cognition in the context of real-world problems. %TI Creative Conceptual Change %AU Ashwin Ram %AV http://www.cc.gatech.edu/faculty/ashwin/projects/creative-conceptual-change.html %AB Creative conceptual change involves (a) the construction of new concepts and of coherent belief systems, or theories, relating these concepts, and (b) the modification and extrapolation of existing concepts and theories in novel situations. Computational models of constructive and extrapolative processes in creative conceptual change specify the functions of conceptual change, the mechanisms or algorithms that achieve these functions, and the knowledge and representations that the mechanisms rely on. Implemented systems include ISAAC, a science fiction story understanding program that carries out conceptual change as it reads about concepts different from its own; and SINS, a robot navigation system that autonomously and progressively constructs representational structures that encapsulate the system's sensorimotor experiences. %TI Design Intelligence %AU Ashok Goel %AV http://www.cc.gatech.edu/aimosaic/faculty/goel/di/ %AB The Design Intelligence (DI) research group, led by Professor Ashok Goel, has the following charter: Investigation of basic artificial intelligence issues, e.g., problem solving and learning, in the context of design; Construction of knowledge/experience-based theories of design, computer-based design, and computer-based design education; Design of intelligent systems capable of adaptation and learning; and Exploration of complex AI issues, e.g., creativity, in the context of design. %TI Evaluation of Case-Based Reasoning %AU Ashwin Ram %AV http://www.cc.gatech.edu/faculty/ashwin/projects/case-based-reasoning.html %AB We are developing a methodology for the computational and empirical analysis of the capabilities and behavior of reasoning systems in order to evaluate the effectiveness and generality of case-based reasoning approaches. Formal computational models facilitate comparisons across different reasoning methods, different classes of problems, and different hardware architectures. For example, they can be used to analyze the effect of the utility problem in case-based reasoning and other machine learning systems. Empirical models enable us to understand the behavior of the system in terms of the theory and design of the computational model, to select the best system configuration for a given domain, to verify that the system improves its performance with experience, and to predict how the system will behave when the characteristics of the domain or problem change. %TI Experience-Based Agency (a.k.a. Case-Based Reasoning: The Next Generation) %AU Ashwin Ram %AV http://www.cc.gatech.edu/faculty/ashwin/projects/case-based-reasoning.html %AB The real world has many properties that present challenges for the design of intelligent agents: it is dynamic, unpredictable, and independent, poses poorly structured problems, and places bounds on the resources available to agents. Agents that opearate in real worlds need a wide range of capabilities to deal with them: memory, situation analysis, situativity, resource-bounded cognition, and opportunism. We are developing a theory of experience-based agency which specifies how an agent with the ability to richly represent and store its experiences could remember those experiences with a context-sensitive, asynchronous memory, incorporate those experiences into its reasoning on demand with integration mechanisms, and usefully direct memory and reasoning through the use of a utility-based metacontroller. Implemented systems include NICOLE, a system that merges multiple plans during the course of case-based adaptation in least-committment planning; and SINS, a multistrategy case-based and reinforcement learning system that continuously constructs and adapts its own system-environment model. %TI Goal-Driven Learning %AU Ashwin Ram %AV http://www.cc.gatech.edu/faculty/ashwin/projects/goal-driven-learning.html %AB The central idea underlying goal-driven learning is that because the value of learning depends on how well the learning contributes to achieving the learner's goals, the learning process should be guided by reasoning about the information that is needed to serve those goals. The effectiveness of goal-driven learning depends on being able to make good decisions about when and what to learn, on selecting appropriate strategies for achieving the desired learning, and on guiding the application of the chosen strategies. Implemented systems include Meta-AQUA, a system that learns by reading natural language stories; Meta-TS, a system that learns to troubleshoot in a real-world domain of electronic assembly manufacturing; and AICC, an attribute-incremental concept formation system. %TI Learning and Adaptation in Autonomous Intelligent Agents %AU Ashwin Ram, Ron Arkin %AV http://www.cc.gatech.edu/faculty/ashwin/projects/robot-learning.html %AB This set of projects is investigating two fundamental types of learning mechanisms in intelligent robotics systems: (i) on-line adaptive learning, which allows a system to respond to unexpected situations and learn while engaged in the problem-solving process, and (ii) off-line learning, which allows a system to learn through slower, more intensive learning processes. Learning is carried out using multiple learning methods, including case-based reasoning, reinforcement learning, and genetic algorithms. Of particular concern is the development of methods that allow a reactive control system to improve its performance by learning to tune its behavor to different environments. This requires the system to learn a model of not just the environment but the interaction between the system and the environment in a continuous, on-line manner. However, to ensure that the system does not get bogged down in extensive high-level reasoning, the knowledge represented in the model must be based on perceptual and motor information easily available at the reactive level. Implemented systems include ACBARR, a case-based reactive navigation system; SINS, a multistrategy case-based and reinforcement learning system that continuously constructs and adapts its own system-environment model; and GA-ROBOT, a genetic algorithm for evolving reactive control systems. %TI Natural Language Understanding %AU Ashwin Ram %AV http://www.cc.gatech.edu/faculty/ashwin/projects/natural-language-understanding.html %AB Reading and understanding real texts requires a large range of tasks, including sentence processing, story structure understanding, episodic understanding, explanation, memory, interest management, learning, and so on. We are developing a functional theory of reading which models the complete set of tasks which a reader must perform during the comprehension process. The various tasks maintain a close interaction, exchanging information as needed; this integrated approach lessens the burden on any one task. We are particularly interested in a kind of reading we call creative reading, in which the reader must learn enough about a novel situation, in a short text, in order to accept it as the background for the story, and simultaneously must understand the story itself (for example, consider reading a science fiction story). Implemented systems include AQUA, a system that learns about terrorism by reading newspaper stories about unusual terrorist incidents; and ISAAC, a system that reads short science fiction stories. %TI Natural Language and Reasoning %AU Susan Bovair, Jeffrey Donnell, Kurt Eiselt, Masato Kikuchi, Wendy Newstetter, Ashwin Ram %AV http://www.cc.gatech.edu/faculty/ashwin/projects/nlr/ %AB The Natural Language and Reasoning (NLR) group at Georgia Tech is an interdisciplinary team of faculty and students in Artificial Intelligence, Cognitive Psychology, and Language. We are interested in studying language understanding "in the large", that is, in the context of a real task. This requires us to examine a range of issues relevant to sentence understanding ranging from syntactic and semantic processing of individual sentences to explanation and learning processes involved in understanding how sentences fit together in real text. %TI Research in Creativity %AU Ashok Goel, Mark Guzdial, Janet Kolodner, Nancy Nersessian, Ashwin Ram, Linda Wills %AV http://www.cc.gatech.edu/aimosaic/faculty/kolodner/creativity/ %AB We have been studying creative reasoning in several different domains, with a goal of producing computational process models of creativity. This will have implications for design education and suggest ways of supporting and enhancing the creativity of people. %TI The ADAM Project %AU Ethan Beisher %AV http://www.prism.gatech.edu/~eb79/adam %AB ADAM stands for Assembly Description / Assembly Modification. It is a manufacturing / Mechanical Engineering system designed to explore the knowledge of assembly and disassembly processes. %TI The AUTOGNOSTIC Project %AU Ashok Goel %AV http://www.cc.gatech.edu/aimosaic/faculty/goel/di/learning.html %AB AUTOGNOSTIC is a computational model of model-based self-redesign. The AUTOGNOSTIC project explores the issues of functional and strategic self-redesign in AI systems. The key idea is that AI systems, such as ROUTER, can be viewed as abstract devices and failure-driven learning can be viewed as device redesign problem solving. A meta-model of how the problem solving in the AI system works can enable the system to reflect on its experiences, especially its failures, and to redesign the problem solver. %TI The IDeAL Project %AU Ashok Goel %AV http://www.cc.gatech.edu/aimosaic/faculty/goel/di/learning.html %AB Mental models are schemes for representing and organizing knowledge and analogy is a reasoning strategy. Both play central roles in reasoning and learning. The IDeAL project explores interactions between device models and analogy in the context of innovative design. IDeAL is a computational theory that provides an account of both how models enable analogical transfer and how analogies enable the construction of models. It uses two kinds of models: structure-behavior-function models of specific devices, and behavior-function models of generic physical processes and engineering mechanisms. The generic BF models are learned by abstraction from device-specific SBF models and mediate cross-domain analogical transfer. SBF models of new devices are acquired in part by adapting the models of similar devices and partly through analogical transfer in the form of BF models. Current work focuses on using IDeAL's computational theory of model-based analogy for understanding specific examples of scientific discovery in which both mental models and analogy apparently played a critical role. %TI The KRITIK Project: Adaptive Design %AU Ashok Goel %AV http://www.cc.gatech.edu/aimosaic/faculty/goel/di/conceptual-design.html %AB Much of everyday design appears to be adaptive, i.e., new designs apparently are created by adapting past design cases. KRITIK is a computational model of adaptive design. The model integrates case-based and model-based reasoning: while the high-level process is case-based, the qualitative structure-behavior-function (SBF) device models associated with the design cases provide the vocabulary and the strategies for design retrieval, adaptation, evaluation and storage. Current work focuses on developing an interactive multimedia design and learning environment called Interactive KRITIK. %TI The RAURA Project %AU Ashok Goel %AV http://www.cc.gatech.edu/aimosaic/faculty/goel/di/multi-strategy.html %AB RAURA is an instantiation of ROUTER in the AuRA architecture (AuRA is a hybrid architecture for autonomous mobile robots capable of both deliberative planning and reactive control developed by the Robotics group). %TI The REFLECS Project %AU Ashok Goel %AV http://www.cc.gatech.edu/aimosaic/faculty/goel/di/multi-strategy.html %AB REFLECS is a reflective system capable of learning strategies for reactive control by diagnosing and repairing failed reactive strategies. %TI The ROUTER Project: Multi-Strategy Systems %AU Ashok Goel %AV http://www.cc.gatech.edu/aimosaic/faculty/goel/di/multi-strategy.html %AB Router is an integrated planning and learning system for autonomous mobile robots. It integrates case-based and model-based reasoning for planning navigation paths in geographical spaces. In the model-based mode, the system plans paths by heuristically searching a hierarchical model of the navigation space; and in the case-based mode, it plans new paths by adapting and combining previously planned paths. It uses a form of meta-reasoning for selecting a specific method at run-time. It also chunks the new plans into cases for potential reuse. %TI The Science Education Advisor %AU Janet Kolodner and Terry Chandler %AV http://www.cc.gatech.edu/aimosaic/faculty/kolodner/AI&ED.html %AB The Science Education Advisor (SCI-ED) is a case-based hypermedia browsing tool designed to collect and dispenses ideas and advice for the teaching of science in elementary school. Our approach has been to apply what the researchers in the cognitive and educational communities have learned about cognitive functioning to build a system which augments the memories of teachers, advises teachers on how to best apply a activity or pedagogical approach to their class, and, promotes communication and collaboration within and across the educational and scientific community.