|
|||||||||
| NYU Psychology | Programs | Courses | Research | Faculty | People | Events | Contacts | [Internal] |
| Todd Gureckis | |
||||||
| Research | Biography | Publications | Address | ||||
|
Assistant Professor of Psychology My research interests center on the memory, learning, and decision processes which allow us to carry out intelligent and adaptive behaviors. I am particularly interested in how we uncover important and useful regularities about the environment. For example, how do we all come to agree on a similar idea of the concept "mammal"? How do our learning experiences shape our perception of the world around us? How does our ability to quickly learn about and adapt to our local environment help us learn new skills and behaviors? The goal of my research is to enrich our understanding of the mechanisms which support these diverse behaviors, how they develop and change over the course of our lives, and how they are impacted by disease or brain damage. Central to my work is the use of computational models as a tool for integrating and directing research. Computational models are psychological theories which have been specified in enough detail to be run as computer programs. Computational modeling provides a powerful scientific framework for evaluating theories of cognitive function. For example, models can help organize diverse sets of findings which might seem otherwise unrelated, and predictions derived from competing models can be used to guide empirical research. In addition, insights from cognitive models can inform the development of artificial systems capable of learning on their own. Within this broad theme, my research has focused on a number of specific areas. The first deals with category and concept learning or how people learn to generalize from the instances they see in the world. For example, how do young children learn which plants belong to the category "tree" and which are "flowers"? Category learning is a critical cognitive ability which underlies a vast array of abilities such as object recognition and language acquisition. My research in this area seeks a better understanding of the learning processes which support this ability and how they might be organized or implemented in the brain. My work in this area combines a variety of methodologies including behavioral studies, neuro-imaging, and computational modeling. Obviously, categories are just one type of information we need to know in order successfully interact with the world. A second, but related line of work deals with the more general phenomenon of statistical learning or our ability to learn and exploit probabilistic regularities in a complex and noisy environment. In recent years, there has been a growing recognition that the world provides us with considerable structure in the form of statistical information which our brains seem to effortlessly "data-mine" for sophisticated patterns. For example, infants can use the likelihood that particular speech sounds follow one after another in order to discover the boundaries between words (Saffran, Aslin, & Newport, 1996). My work in this area has attempted to identify the constraints governing this ability and to characterize the range of statistical complexity humans can readily utilize in sequential tasks. My research also considers how people learn through direct interaction with their environment. Imagine a young child just beginning to learn about how the world works. Some of the time, an adult is nearby to provide feedback or instruction, but often times the child must learn by doing or through trial and error. In some of my recent work, I've looked at how people learn effective behavioral strategies through interacting with dynamic task environments where reward structure continually changes in response to the actions of the individual. In contrast to approaches that treat learners as passive observers, my goal in this work is to develop theories of learning that can inform the active way in which we search, sample, and explore our environment. On the computational side, this work applies the framework of reinforcement learning (Sutton & Barto, 1998) to understand higher-level cognitive behavior and decision making. Reinforcement learning is an agent-based approach to learning through interaction with the environment in pursuit of goal-directed behavior. Outside of psychology, reinforcement learning has been successful in numerous practical applications (such as flying helicopters, Bagnell & Schneider, 2001, and teaching computers to play backgammon, Tesauro, 1994), and in the modeling of biological systems (such as modeling the response of dopamine neurons). However, reinforcement learning is also a promising theoretical approach to studying higher-level human learning since it emphasizes the role of a situated agent interacting with their environment. For example, RL models specify how learners should balance exploration versus exploitation of resources in their environment, how they take into account delayed rewards while learning, and how to assign credit to actions that later lead to successful outcomes.
Finally, many diverse natural systems ranging from ant colonies to the human brain share a remarkable propensity for self-organization, coordination, fault-tolerance, and emergent behavior. A key property that many of these system seem to share is that they are composed in interdependent systems of simple elements that, by virtue of their interaction, give rise to emergent behaviors that cannot be understood by studying the elements themselves in isolation. For example, neurons in our brain give rise to complex thought patterns that are beyond any individual element in the system. In the physical and biological sciences, systems that exhibit these properties are referred to a complex adaptive systems. I have a line of interdisciplinary research that investigates how interacting groups of individuals following their own, self-motivated rules can organize into adaptive groups with emergent properties. Note: I will be starting up my lab in 2008 and interested graduate students are encouraged to contact me.Biography I received a B.S. in Computer/Electrical Engineering (2001) from the University of Texas at Austin. Following the collapse of the tech industry in 2001, I decided to pursue my interests in human and machine learning by enrolling in graduate school in the psychology department at UT Austin. I received a M.A. and Ph.D. (2005) in psychology from UT Austin under Brad Love. After completing my Ph.D., I was employed as a postdoctoral research associate in the Psychological and Brain Sciences Department at Indiana University where I worked with a number of faculty including Rob Goldstone, Rob Nosofsky, and Peter Todd. I start as an Assistant Professor in the NYU Psychology Department in Jan. 2008. Goldstone, R.L. and Roberts, M.E. and Gureckis, T.M. (in press) Emergent Processes in Group Behavior. Current Directions in Psychological Science [pdf] Gureckis, T.M. and Love, B.C. (2007) Behaviorism Reborn? Statistical Learning as Simple Conditioning. Proceedings of the 28th Annual Conference of Cognitive Science Society. [pdf] Gureckis, T.M. and Goldstone, R.L. (2007) Schema. The Cambridge Encyclopedia. Love, B.C. and Gureckis, T.M. (2007). Models in Search of the Brain. Cognitive and Affective Behavioral Neuroscience [pdf] Goldstone, R.L., Roberts, M., Mason, W. and Gureckis, T.M. (2007). Collective Search in Concrete and Abstract Spaces. Decision modeling and behavior in uncertain and complex environments. Kugler, T., Smith, C., and Connelly, T. (Eds.). Springer Press. Gureckis, T.M. and Love, B.C. (2006). Bridging Levels: Using a Cognitive Model to Connect Brain and Behavior in Category Learning Proceedings of the 28th Annual Conference of Cognitive Science Society. [pdf] Gureckis, T.M. and Goldstone, R.L. (2006). Thinking in Groups. Pragmatics and Cognition. [pdf] Gureckis, T.M. and Love, B.C. (2005). A Critical Look at the Mechanisms Underlying Implicit Sequence Learning. Proceedings of the 27th Annual Conference of Cognitive Science Society. [pdf] Love, B.C. and Gureckis, T.M. (2005). Modeling Learning Under the Influence of Culture. In Categorization inside and outside the lab: Festschrift in honor of Douglas L. Medin Edited by Ahn, W., Goldstone, R., Markman, A., Wolff, P. and Love, B. Washington D.C., APA Publisher. Love, B.C. and Gureckis, T.M. (2004). The Hippocampus: Where a Cognitive Model meets Cognitive Neuroscience. Proceedings of the 26th Annual Conference of Cognitive Science Society. [pdf] Gureckis, T.M. and Love, B.C. (2004). Common Mechanisms in Infant and Adult Category Learning. Infancy, vol 5, no.2, 173-198. [pdf] Love, B.C., Medin, D.L., and Gureckis, T.M. (2004) SUSTAIN: A Network Model of Category Learning. Psychological Review, 11, 309-332 [pdf] Gureckis, T.M. and Love, B.C. (2003). Human Unsupervised and Supervised Learning as a Quantitative Distinction. International Journal of Pattern Recognition and Artificial Intelligence, vol. 17, no. 5, 885-901. [pdf] Gureckis, T.M. and Love, B.C. (2003). Towards a Unified Account of Supervised and Unsupervised Learning. Journal of Experimental and Theoretical Artifical Intelligence, 15, 1-24. [pdf]
|
