Dr. Ron Sun

Ron Sun, Ph.D.  

Professor
Cognitive Science Department 
Rensselaer Polytechnic Institute 
110 Eighth Street, Carnegie 302A  
Troy, New York 12180, USA  

Phone: (518) 276 3409
Homepage: external homwpage
Email: rsun [at] rpi.edu

Index


INTERESTS

Dr. Sun has been instrumental in organizing some of the most important events concerning hybrid systems (mostly in relation to cognitive modeling), such as (co)chairing the 1992 AAAI Workshop on Integrating Connectionist and Symbolic Processes, the 1995 IJCAI Workshop on Connectionist Symbolic Integration, the 1996 AAAI Workshop on Computational Cognitive Modeling, the 1998 NIPS Workshop on Hybrid Connectionist Symbolic Systems, the 1999 IJCAI Workshop on Sequence Learning, the 2001 CogSci Symposium on Implicit and Explicit Cognition, the 2001 ICCS Symposium on Cognitive Modeling, and the 2003 IJCAI Workshop on Cognitive Modeling and Multi-Agent Simulation, the 2006 AAAI Workshop on Cognitive Modeling and Social Simulation, the 2006 CogSci Symposium on Implicit and Explicit Learning, and (co)editing the 1994 Connection Science special issue on hybrid models, the 1997 IEEE Transactions on Neural Networks special issue on hybrid models, and the 2001 Cognitive Systems Research special issue on multi-agent learning.

He was the program chair of IJCNN 2007 held in Orlando, Florida. He was the general chair and the program chair of the 2006 Cognitive Science Society Conference held in Vancouver, Canada. He was the program co-chair of the 2005 WI-IAT Conference in Compiegne, France. He has also been on the program committees of many national and international conferences, such as CogSci (2002, 2003, 2005, 2006), ICCM (2001, 2003, 2004, 2006, 2007), TSC (2002), ASSC (2003), AAAI (1993, 1997, 1999, 2006), IJCNN (1999, 2000, 2002, 2007), ICONIP (1997, 1999, 2001, 2004, 2006), AAMAS (2005), IAT (1999, 2001, 2003, 2004, 2005), and PRIMA (2000, 2001, 2002, 2003, 2004, 2005).

He has published more than 150 papers and 7 books in this area. He has been an invited, plenary, or keynote speaker at many conferences: the First New Zealand International Two-Stream Conference on Artificial Neural Networks and Expert Systems (ANNES'93, Dunedin, New Zealand), the International Symposium on Expert Knowledge and Neural Heuristics (Pensacola, Florida; 1994), the Symposium on Autonomous Robots (Ulm, Germany; 1997), the Midwest Conference on Artificial Intelligence and Cognitive Science (Fayetteville, Arkansas; 2000), the International Conference on Neural Information Processing (Shanghai, China; 2001), Erice 2002 (Erice, Sicily, Italy; 2002), the Symposium on Cognitive Architectures (Stanford, California; 2003), the Joint Symposium of SAIS/SSLS (Orebro, Sweden; 2003), MICS 2005 (Saratoga Spring, New York; 2005), the Workshop on Neural-Symbolic Learning and Reasoning (Edinburgh, Scotland; 2005), the 9th Knowledge-Based Intelligent Information and Engineering Systems Conference (Melbourne, Australia; 2005), PRIMA 2005 (Kuala Lumpur, Malaysia; 2005), The Conference on "To Think and Act like a Scientist: The Roles of Inquiry, Research, and Technology" (Lubbock, Texas; 2006), the Workshop on Model Comparison and Model Validation (Syracuse, New York; 2006), the NIAS Workshop on Minds in Interaction at the Netherlands Institute for Advanced Study in the Humanities and Social Sciences (Wassenaar, Netherlands; 2006), the Workshop on "Combining Cognitive Plausibility with Social Realism: Promises and Pitfalls of Multi-Agent Simulation" (University of Groningen, Groningen, Netherlands; 2006), the WICI International Workshop on "Web Intelligence Meets Brain Informatics" (Beijing, China; 2006), the Mind Forum (Helsinki, Finland; March 2008), the International Conference on Adaptive Knowledge Representation and Reasoning (Helsinki, Finland; September 2008), AAAI Fall Symposium 2009 Multi-Representational Architectures for Human-Level Intelligence (Washington DC; November 5-7, 2009), The 16th International Conference on Neural Information Processing (ICONIP 2009) (Bangkok, Thailand; December 1-5, 2009), as well as the special sessions of IEEE-WCCI'94, IIZUKA'94, IEEE-ICNN'96, ICONIP'97, ANNES'97, IJCNN'98, IEEE-FUZZY'98, IJCNN'99, IJCNN'00, IJCNN'02, and so on.

Dr. Sun is the founding co-editor-in-chief of the journal Cognitive Systems Research (Elsevier), and also serves on the editorial boards of Connection Science, Cognitive Computation, Neural Information Processing--Letters and Reviews, International Journal of Hybrid Intelligent Systems, and so on. He is on the Governing Board of the Cognitive Science Society, and the Board of Governors of the International Neural Network Society. He received the 1991 David Marr Award from Cognitive Science Society (at the Thirteenth Annual Conference of Cognitive Science Society), and will receive the 2008 Hebb Award from the International Neural Networks Society. He is a senior member of IEEE. He is listed in Marquis Who's Who in America (the 53rd, 56th, and 57th edition), Marquis Who's Who in the World (the 16th, 18th, 20th, and 24th edition), and Marquis Who's Who in Science and Engineering (the 4th, 5th, 6th, 7th, 8th, and 9th edition).  


ANNOUNCEMENTS

1. To see a description of his recent books, click on a title:

2. To see a description of the recent conferences, workshops, and/or journal special issues he (co)organized, click on a title:

3. To obtain a description of, and/or to access, the HYBRID LIST (an electronic mailing list devoted to hybrid systems of various sorts, involving connectionist, symbolic, evolutionary, and fuzzy models, moderated by Ron Sun), click here.

4. To see a brief description of our Ph.D program in Cognitive Science, AI and Neural Nets, click here. If you need application forms for the Ph.D program, click here.

Note that I prefer only to supervise graduate students and post-docs who know a lot about my research (i.e., read some of my papers; see below) and wish to do related work. However, I am willing to consider proposals that are slightly afield from exceptionally outstanding students.





ON-LINE COGNITIVE SCIENCE AND AI RESOURCES


1. * SOME JOURNALS (of which Ron Sun serves on the editorial board):

Cognitive Systems Research, published by Elsevier


Neural Networks

Connection Science

International Journal of Hybrid Intelligent Systems

Neural Information Processing--Letters and Reviews

Cognitive Computation

2. * HYBRID SYSTEMS RESOURCES:

The Hybrid Systems Resources Page

including:

An introduction to hybrid systems (by Ron Sun, an entry in International Encyclopedia of Social and Behavioral Sciences)

Surveys of hybrid systems: An article by Ron Sun (appeared in: Connectionist-Symbolic Integration. Lawrence Erlbaum Associates. 1997), an article by S. Wermter and R. Sun (appeared in: S. Wermter and R. Sun, eds. Hybrid Neural Systems. Springer-Verlag, Heidelberg. 2000) (PDF), an article by A. Browne and R. Sun (appeared in: Expert Systems, Vol.16, No.3, 189-207. 1999), and another article by A. Browne and R. Sun (appeared in: Neural Networks, 2001).

Hybrid List moderated by Ron Sun

The 1994 Bibliography on Connectionist Symbolic Integration (edited by Ron Sun, appeared in the book Computational Architectures Integrating Symbolic and Connectionist Processing, published by Kluwer).

3. * COGNITIVE ARCHITECTURES RESOURCES


4. * OTHER RESOURCES



RESEARCH DESCRIPTION

My research interest lies in the study and modeling of cognitive agents, especially in their abilities to learn, reason, and act in the real world. More specifically, my research can be categorized into the following areas: human and machine learning, connectionist reasoning and knowledge representation, hybrid models, as well as multi-agent interaction and cognitive social simulation.

Human and Machine Learning

My research in this area is concerned with skill learning in various domains, ranging from highly intellectual to sensory-motor tasks. The goal is two fold: to better understand human skill learning in various domains and to develop more unified learning models (cognitive architectures) for skill learning tasks. This work thus includes both psychological experiments/data collection and computational simulation and model development. See learning.

A hybrid connectionist model Clarion has been developed, which combines both procedural knowledge and declarative knowledge in one framework. Learning in this architecture is accomplished by reinforcement learning supplemented with rule induction, so that the resulting model is parsimonious in structure and possesses a variety of reasoning and decision-making capabilities. The model performs autonomous learning. It develops different types of representations, symbolic and subsymbolic, simultaneously along side each other. The model is being used to simulate a variety of human skill learning data, including a navigation task, a dynamic control task, a serial reaction time task, and artificial grammar learning, and it starts to shed new light on human learning. I intend to continue this line of exploration for a long time.

The cognitive modeling work leads to the study of some important machine learning techniques, which are interesting in their own right. One of the focuses is the extraction of explicit plans (open-loop policies) based on the results of reinforcement learning, to enable explicit reasoning of plans, without a priori domain knowledge to begin with. A variety of algorithms will be explored for such plan extraction.

Yet another focus is the development of modular reinforcement learning models, in which multiple modules (or agents) compete and cooperate with each other to accomplish tasks, without a priori division of the tasks (i.e., without using any a priori domain-specific knowledge). Such models are useful for cognitive modeling of certain human learning situations as well as in engineering applications. See multi-agent learning.

This work has been supported by major grants from Cognitive Science Program, Office Of Naval Research, and Army Research Institute.

Connectionist Reasoning and Knowledge Representation

For the past several years, my research was mainly concerned with everyday commonsense reasoning by agents. This type of reasoning was characterized by a mixture of rule-based and similarity-based processes, exhibiting both rigor and flexibility (as demonstrated in my AIJ paper). To capture such reasoning, I developed a hybrid connectionist architecture (named CONSYDERR) with both localist and distributed components, that unified rule-based and similarity-based processes and accounted for a variety of CSR patterns. See reasoning.

Within the framework, the following issues were also investigated: (1) The connectionist implementations of rules, logics, and schemas, and the variable binding problem in such implementations. They formed the basis for complex reasoning in connectionist models. (2) Inheritance reasoning, which is an integral part of many CSR patterns. Within CONSYDERR, an intensional approach was developed that works in constant time. This work suggests that other similar reasoning patterns may also be handled intensionally. (3) Causality, which is an important commonsense construct. A connectionist account was developed based on CONSYDERR, which extended the existing logic-based account and dealt better with the inexact, cumulative, and subjective nature of commonsense causal reasoning.

Some attempts have also been made to extend the framework to deal with metaphor and analogy. Further work will be done to refine the architecture and to account for human CSR quantitatively.

Hybrid Models

The above two categories of work lead to the development of two major hybrid models: CONSYDERR and CLARION (and their numerous variations and implementations). My interest in hybrid models lies mainly in developing more powerful, more integrated models that are capable of autonomous on-line learning, acquiring both symbolic and subsymbolic knowledge and utilizing their synergy to achieve better performance.

To do so, I have been exploring, and will continue to explore, psychological models of human learning, especially those concerned with integrated learning of multiple forms of knowledge, as well as machine learning and neural network techniques and theories, especially those concerned with reinforcement learning. I hope a synthesis of these two strands of work will lead to significant advances in developing hybrid models by providing new insights and impetus.



SELECTED PUBLICATIONS


To download the PostScript versions of the papers, try also the ftp site. Click here.




Click for Albany, NY Forecast Click for Albany, NY Forecast



RPI Cognitive Science Dept Homepage