| Name | Department | Research Interests | |
|---|---|---|---|
| Adrian Kujaneck Agogino | agogino@ece.utexas.edu | Computer Engineering | I am interested in created various types of intelligent beings (sometimes
called intelligent agents). One area I find interesting is intelligent
artificial "actors", which in some ways try to act like real people or
animals, or at least try to act plausibly like something living. These
actors could be used in virtual reality settings, games or something more
mundane like the Microsoft Office animated paperclip. I am also interested in intelligent control of real devices like robots. In some ways this is similar to virtual agents except that you have to be more careful about what you let the robot do. Robots also need processing of real world inputs, like vision processing. I would be interested in collaborating on a project with virtual intelligent beings, robots or vision. I have lots of experience with neural networks and some in AI and computer graphics. I generally think of these problems in terms of neural networks and program them up in C or Java. It would be great if I could find a partner with knowledge of animal/human behavior or robotics and visual perception. |
| April Bleske | bleske@mail.utexas.edu | Psychology | I study human relationships from an evolutionary perspective. Much
of my research focuses on opposite-sex friendship, same-sex friendship,
sexual strategies, jealousy, and mate poaching (i.e.,stealing). In general,
I use Sexual Selection and Parental Investment Theory to generate
hypotheses about sex differences in humans' evolved psychologies. I use a mixture of methods. In much of my research I use self-report. For example, the findings from several survey studies I have conducted suggest that, compared to women, men desire sex with their opposite-sex friends more often and initiate friendships for that reason more often. Compared to men, women prefer friends who are physically strong and judge the lack of protection as a more important reason for dissolving an opposite-sex friendship. These results support the hypothesis that opposite-sex friendship is an evolved strategy by which men gain short-term sex, and women gain protection. Self-report data is problematic, however, since evolutionary psychologists don't expect humans to be consciously aware of their evolved preferences, desires, decision-making tactics, etc. The findings above support hypotheses about evolved strategies, but are not conclusive by any means. So, some of my reserach is experimental. This semester, for example, I am bringing couples into the lab and exposing them to a confederate. The confederate's job is to flirt with one member of the couple, thus evoking jealousy in the other (we measure jealousy by self-report as well as by behavior). We predict that for women, promiscuously dressed rivals will evoke more jealousy than will non-promiscuously dressed rivals, and that for men, high-status rivals will be more threatening than low-status rivals. These predictions were generated by sexual strategies theory.
|
| Darren Cambridge | d.cambridge@mail.utexas.edu | My research examines distributed cognition in large electronic information spaces. I am interested in how networks of humans and machines "thick" collectively in order to learn. "To learn" in this context means to generate, store and apply knowledge shared by the social group. Specifically, I am interested in how communities of practice within these spaces overlap at their peripheries and cross-appropriate concepts, tools, and practices to either form new communities of practice or larger constellations of communities of practice that transverse the original communities. From the results of this inquiry, I'd like learn how to better design electronic collaboration environments to encourage learning (taken in this broad, social sense) within heterogeneous user populations. I am especially interested in ways of using artificial intelligence techniques, such as collaborative filtering, to help increase the rate of learning.
This line of inquiry requires looking at very complex systems that are virtually impossible to simulate with a computational model or to reproduce in a controlled setting. One has to examine systems "in the wild." The method I've taken is to actually build such a space to support a real life collection of communities of practice likely to engage in cross-appropriation -- the 125 groups of faculty and administrators on the campuses of a wide variety of higher education institutions across the country involved in the American Association for Higher Education (AAHE) Campus Program, part of the Carnegie Academy for the Scholarship of Teaching and Learning (CASTL). The goal of CASTL is to promote the practice and valuation of the scholarship of teaching and learning in higher education. The campus groups are currently engaged in local conversation in which they are determining what the scholarship of teaching and learning means for them in their local context and how they might work to cultivate it together. In the second stage of the program, individuals and campus groups will engage in conversations and resources sharing nationally. My project, the WebCenter (lame name, I know), is designed to support this second phase. A basic prototype is now up at For this class, I would like plan a project related to the larger project of the WebCenter. We have fairly good funding now, and are very likely a large amount of additional grant money over the course of the year, so there's a good chance that the project for this class could evolve into a substantial, longer-term research opportunity. There are two types of projects that I think would be quite interesting and productive. I need to draw on different types of expertise for both. First, we are in process of developing an automated peer-review system for the site that collects numerical evaluations from reviewers using a particular rubric and calculates whether or not their in consensus amongst the group based on some fairly sophisticated (at least for me) statistical analysis. This system has been worked out theoretically, but it has not been built so that its effectiveness can be evaluated. I could *really* use a collaborator who has strong mathematics and programming skills to help figure out how to make the theory into reality. (For more details on this system, see the draft of my Computer Supported Collaborative Learning 99 paper posted at http://aahe.ital.utexas.edu/CSCL.doc>http://aahe.ital.utexas.edu/CSCL.doc) Second, once the collaborative filtering system (and, eventually, the peer review system) are integrated into the site and Campus Program participants have used it for a few months, we need to design a research project that evaluates the effectiveness of the system in stimulating cross-appropriations, the formation of new communities of practice, and/or a global consensus about what the scholarship of teaching and learning is and how best to do it. Such a study would likely draw in both the data collected by the server logs, the information and documents stored on the system, and surveys collected from or interviews conducted with participants. A collaborator with experience in social science methodologies and the imagination to combine and adapt them to this project would be an immense help. I am also interested in the application of intelligent agents to teaching and learning in general, and I would be interested in anyone else's project ideas that touch on this subject. |
|
| Chun-Chi Jonathan Chen | ccchen@cs.utexas.edu | Computer Science | Mainly I am interested on how humans perceive speech in a continuous
manner, how each syllable is processed inside of the brain and how
machines can model this process in a discrete, digital way. I am
especially interested in exploring an infant's ability to produce sound
and learn to speak. I would like to build a computational model to carry out the hypothesis, if any. |
| Miranda Falk | zinj@mail.utexas.edu | Linguistics | My background is primarily one of psycholinguistics. My studies have
encompassed first and second language acquisition, neurolinguistics,
neurobiology, sociolinguistic perception. With respect to this class, I am really interested in memory retention and consolidation. Since I probably won't get much funding towards lesioning people, perhaps a more benign study would suffice. Recently, I've been toying with the idea of trying to analyze the effects of sleep on memory consolidation. I am not at all fixated on this topic, however; I would be open to a wide range of possibilities. |
| David Han | dhan@lips.utexas.edu | Electrical and Computer Engineering | My research concentrates mainly on artificial intelligence planning
processes and the integration of these processes into multi-agent
systems. This includes coordination of interagent behavior. In our
research group, we have ~6 students working together to implement
reasoning algorithms in a Multi-Agent Testbed and simulation to test
agent behavior under various conditions. I am interested in learning more about how (conscious) decisions are made to help my understanding of (deliberative and reactive) planning processes on the computer. I have a good programming backgroung in Lisp, Java. and C++.
|
| Malcom Gibran Haynes | mghaynes@cs.utexas.edu | Computer Science | I am a ship without a sail. I'm a Computer Science major, but I have
basically worked on networks and distributed computation. My AI
experience is limited to one overview course that introduced AI
heuristics, neural networks, and expert systems. I pretty much find all
the topics we have discussed thus far in class very interesting, so it's
hard to say what I would like to do for a project. Basically, if anyone
thinks they may have a use for an CS person in a project I'm the guy. Perhaps the topic in class I have found most interesting in vision and human perception. It would be interesting to work on some project that looks at how human minds perceive reality vs. the actual reality. (I'm thinking of experiments like lighting up the green and red light and human perceives it as light moving left and changing color) But, as stated earlier, I would be happy to work on most anything. The whole field is interesting. |
| Christopher Johnson | cpjohnson@mail.utexas.edu | Journalism | I am interested in Mark Johnson's work on metaphor as an extension of basic cognitive "image schemata" that infants may learn from their first few months of acting on and observing actions in the physical world. Johnson and Lakoff have argued that metaphor is the foundation of all language, and Johnson has gone on to argue that what he sees as the inherent physicality of all metaphorical structures reveals a physical basis for all language and thought. Human understanding, he suggests, comes first from understanding the physical world and the body. I am interested in whether Johnson's proposed basic "image schemata"--simple physical metaphors such as the "container" metaphor (one object can be contained within another), the "blockage" metaphor (one object blocks the trajectory of another object), or the "compulsion" metaphor (one object pushes another object into motion)--could be computer modeled as basic structural foundations for perception and understanding of the physical world for AI (I don't know what things like this may have been done in AI at this point). This might be rather ambitious, but the idea would explore whether computers could begin to analogize between similar kinds of physical (perceptual) experiences. Ultimately, I would like to know if this could at all be a likely approach to testing Johnson's idea that human cognition, even at its most abstract levels, is an elaborate extension of certain basic physical paradigms for all possible actions in and on the world. Since my background is in liberal arts--literature, communication, and some psychology, I would likely need to collaborate with someone (or two or three) who has some background in computer modelling and takes an interest in language. So far, my methods in researching this topic have involved looking at any research that suggests any physical, bodily determination in human cognition and language, such as studies on gesturing in the congenitally blind. |
| Nicole Kime | nmkime@mail.utexas.edu | Zoology | I am interested in animal communication, especially in the context of
sexual selection. In many species, females (or, less often, males)
select among potential mates on the basis of male secondary sexual
characteristics. In our lab, we study the evolution and maintenence of
these characters and mating preferences. I mostly work with frogs, and
do both computer analysis of male vocalizations and behavioral
experiments with females to assess their responses to variation in male
calls. I also have some limited experience with neuroscience, am working
to put together a new neurophysiology lab, and expect to start looking
at auditory responses to natural acoustic signals as soon as I can get
it all to work! With resepct to this course, I should mention that I've been increasingly interested in the cognitive constraints that can limit the evolution of behavioral repertoires. This has obvious implications for my own work on communication, as well as many other aspects of behavior. I've been thinking about these sorts of questions quite a bit lately, and would really like to look at some of them further, using either computer simulations or empirical studies. |
| Mark Mallon | mw.mallon@mail.utexas.edu | Psychology | Currently I'm involved with research exploring the perception of rhythm.
I am also motivated to study the effects of stress (be it physical or
mental) upon cognition (specifically, performance tasks).
I have as well other psychophysic interests including stereopsis, with
which I have some lab experience. In both that lab and my current domicile, the most prominent method employed is that of writing a Matlab program which, once stimuli are presented to a subject, their response and speed of such is analyzed to ascertain what the most likely kind of process the brain is using for said task. Neuroscience methods are also familiar to me, as I spent a few years (in a sleep behavior laboratory) implanting, running, and getting the histology on rats. There we were looking for an animal model of Post Traumatic Stress Disorder, via ablation or neurotransmitter infusion. Memory (all types) interests me too.
|
| Alisa Marzilli | alisa@ticam.utexas.edu | Computational and Applied Math | My research thus far has involved robotic spatial perception. I have been working with a robot wheelchair that uses laser range scanners to detect obstacles in its view. It takes the data from these scanners and builds a "map" of its environment by building a grid of probabilities; each probability corresponds to "how occupied" a region of the environment is. I generally "tackle" problems by writing up computer programs (in C or java) and testing the ideas. So, for a group project, I would be interested in something in robotics, as well as in venturing into Natural Language Processing or Machine Learning (but I have minimal experience in the latter fields!). |
| Mary O'Brien | mgrantham@mail.utexas.edu | German | Linguistics of all types, especially first language acquisition and second language learning. I am eapecially interested in phonloogy (and phonetics). Most first language acquisition data involves listening to a child for extended periods of time and transcribing all utterances made by that child. Because this is so time-consuming, I have often used data gathered by others and from that found trends that occur within that one child. (Second language learning data is much easier to gather, although it is often more difficult to analyze. I could use my own German 508K students for the data.) |
| Carlos Pacheco | cpacheco@mail.utexas.edu | Computer Science | I would like to model neurobiological or cognitive processes in the brain using neural nets. I'd like to explore aspects of individual neuron activity which may have been overlooked or oversimplified in current neural net approaches. I'd also be willing to follow the more traditional approach of designing and training a well-known neural net model using a set of behavioral data, testing the model to study its effectiveness in imitating the observed behavior, and possibly make further predictions of the given behavior. My methods of research would obviously be creating a computer model/simulation, and I would like someone who has more knowledge than I do about neuroscience in order to find interesting aspects of the brain to investigate. |
| moo-Kyoung Song | msong@mail.utexas.edu | Music Theory | I am interested in an interdisciplinary area between linguistics and music, especially in the context of semiology of music. In the current stage of music analysis, a hierarchical theory such as a Schenkerian theory or generative theory of music takes up an important position. I'd like to trace generative theory of music of Jackendoff and Lerdahl which is based on Chomsky's theory of transformational grammar and further to illuminate how it influences on the practice of music analysis. I'm also interested in an approach of the informational processing to music perception. I think that music is a language of a specific form in that it communicates something. I am finding co-workers who are interested in linguistics or informational processing, especilly in aural perception. |
| Dan Tecuci | tecuci@cs.utexas.edu | Computer Science | My main area of interest is Artificial Intelligence in general, and
Machine Learning in particular as I consider learning to be the central
characteristic of intelligent systems. As a long-term goal I would like
to study the differences between what a computer can do and what a human
can do (at the "information processing" levels) and the nature of these
differences (maybe found at lower levels - "representation and
algorithm" or "implementation" in Marr's terminology). Possible start
points are: the difference between continuous and discrete processing of
information streams, serial versus massively parallel processing. I have a good experience in computer programming (Lisp, C, C++, Java) and modeling. For my graduation thesis I developed a system that refines an incomplete and partially incorrect knowledge base using machine learning and knowledge elicitation from the interaction with the expert user. For this class I would be interested to collaborate with other student(s) that have more experience in designing and conducting experiments with humans so that we can develop, model and test (implement) a hypothesis concerning aspects human cognition (learning, language, vision, memory, social organization, etc.) |
| J. Aaron Tropp | eclectic@mail.utexas.edu | Computational and Applied Mathematics | Since I am a neophyte graduate student, I don't have any focused research
interest. My primary area of application in CAM (Computational and Applied
Mathematics) is graphics and data visualization. I am also curious about
models of reasoning and about natural language processing. As for my background: I have bachelor's degrees in mathematics and Plan II, UT's liberal arts honors program. I also took a significant number of courses (4+) in the computer sciences, philosophy, history, Latin, French and literature. My primary extracurricular activities involved newspaper journalism and design. As a result, I am not really qualified to do anything. |
| Yeh Hong-ming | hongming@cs.utexas.edu | Computer Science | My research interests include artificial intelligence, especially in natural language understanding, knowledge representation, and commonsense reasoning. In addition to traditional artificial intelligence and machine learning techniques, I am interested in soft computing such as genetic algorithms and neural networks. |
| Jiun-Shiung Wu | jswu@mail.utexas.edu | Linguistics | I am interested how formal semantics is used in knowledge representation. Right now, I am particularly interested in scope ambiguity. If possible, I would like to explore how the context information de-ambiguates scope ambiguity and how to represent it on computer. |