Research in Neural and Statistical Computation
at the University of Colorado, Boulder
The University of Colorado at
Boulder
provides an outstanding interdisciplinary
environment for research and graduate training in Neural and
Statistical
Computation in the fields of Artificial Intelligence, Cognitive
Science,
and Engineering. Our research spans the following topics:
- machine learning
- neural network theory
- reinforcement learning
- applications of machine learning techniques to engineering
problems
- adaptive control of complex, nonlinear systems
- computational models of perception, attention, and cognition
- statistical approaches to natural language understanding
- speech recognition
- mechanisms of learning in the brain
- experimental studies using child and adult behavioral measures,
fMRI, and ERP
- optoelectronic implementations of neural networks
While these research topics are diverse and involve different subsets
of the faculty, faculty interactions lead to many synergies among the
topics. A weekly research meeting, the Boulder Computational
Learning Group, serves to bring together
the interdisciplinary community.
Faculty
- Michael
Brandemuehl
- Adaptive control for building systems; simulation and testing of
energy systems; HVAC systems
- Moncef
Krarti
- Building systems modeling with neural networks and statistical
techniques
- Jan
Kreider
- Applications of neural and statistical computation to building
energy prediction, nonlinear adaptive control problems, and building
system diagnosis
- Elizabeth
Bradley
- Artificial intelligence approaches to understanding and modeling
nonlinear dynamics and chaos; nonlinear control
- Greg Grudic
- Theoretical analysis and practical implementation of machine
learning algorithms intended for very high dimensional state spaces;
reinforcement
learning; nonparametric regression; nonparametric classification
algorithms
in large state spaces; machine-learning based robotics
- Larry Hunter
- Development and application of advanced computational techniques
for
biomedicine, particularly the application of machine learning and
statistical
inference techniques to high-throughput molecular assays. Also,
automated
processing of biomedical texts, inference of metabolic and signaling
pathways,
and neurobiologically and evolutionarily informed computational models
of
cognition.
- James
Martin
- Statistical approaches to machine translation, spoken language
help
systems, and the generation of instructional texts; metaphor
understanding
- Michael
Mozer
- Computational models of human attention, perception, and
cognition; engineering applications of neural networks, including the
control of building
energy systems and speech recognition; neural network algorithms for
temporal pattern processing
- Jane Mulligan
- Stereo vision; Modeling and prediction from depth
sequences; Telepresence; Experimental analysis of robotic tasks;
autonomous navigation; robotic manipulation; integration of
computational vision and manipulation.
- Martha Palmer
- Natural language processing and knowledge representation; computational linguistics; machine translation; annotation
- Wayne
Ward
- Spoken language understanding, conversational speech systems,
summarization, integrating stochastic and rule-based language models
- Timothy Brown
- Applications of neural networks to telecommunications (adaptive
control
in broadband networks, high-speed switching processors, equalization);
neural network learning; recurrent neural network design frameworks
- Howard
Demuth
- Applications of artificial neural networks to problems in
control systems, forestry, and chemistry. Authored and maintains the
neural network
toolbox for MATLAB, and teaches Neural Network Design in the spring.
- Kelvin Wagner
- Applying organizational principles of neural networks in the
brain to artificially constructed adaptive systems made from optical
devices. The neurons are implemented using custom integrated circuits
that incorporate photodetectors and light modulators, while the
adaptive synapses are formed as dynamic holographic interconnection
gratings whose strength grows in proportion to the correlated activity
of the source and destination neurons. Architectures, devices, and
simulations have been
developed for self-aligning multilayer holographic optical learning
systems for the implementation of optical back propagation and optical
competitive learning, which are prototypical supervised and
unsupervised learning algorithms.
- Lise Menn
- Child language acquisition; neurolinguistics; aphasia
- Marie
Banich
- cognitive neuroscience of attention, memory, and executive
function;
human neuropsychology
- Eliana Colunga
- Language development, concept acquisition, statistical learning;
methodologies include developmental studies and computational modeling
- Tim Curran
- Human learning and memory from a cognitive neuroscience
perspective, using
behavioral methods derived from cognitive psychology,
neuropsychological studies of the effects of brain injury, and
neuroimaging methods (PET, fMRI, ERP)
- Matt Jones
- human learning and knowledge representation, with emphases on
categorization, similarity, generalization, relational representations,
and sequential decision making
- Walter
Kintsch
- Psychological theories of comprehension, discourse processing,
and higher reasoning based on statistical approaches, including
parallel
relaxation search and latent semantic analysis, a technique that
derives estimates of relatedness from very large text corpora.
- Thomas
Landauer
- Statistical simulation of large-body semantic knowledge
acquisition from
text corpora and mathematical modeling of discourse comprehension
processes
- Akira
Miyake
- Cognitive neuropsychology;
cognitive modeling of normal and pathological cognition ;
visuospatial thinking and imagery;
working memory and executive function
- Yuko Munakata
- Memory, attention, and controlled processing, assessed through
computational models and studies of behavioral dissociations in
children and adults
- Randy O'Reilly
- Biologically constrained computational modeling of cognitive
phenomena. Currently focusing on interactions between hippocampus,
prefrontal cortex, and posterior association cortex in episodic
memory phenomena, and on developing a biologically plausible yet
computationally powerful model of long-term learning in the neocortex.
Graduate Study
Applications for graduate study and further information about graduate
programs can be obtained from the home page of the relevant
department. Those with an interest in cognitive science and cognitive
neuroscience should visit the Institute of
Cognitive Science home page.
Support is generally available for Ph.D. students in the form of
teaching
and research assistantships.
In addition to providing an exciting research environment,
the Boulder area offers
an exceptional quality of life. Spectacularly situated at the foot of
the
Rockies, Boulder provides a wide variety of extraordinary outdoor
activities
and an average of 330 sunny days per year. Together with Denver, Boulder
also
affords a broad range of cultural opportunities.
Courses
This is a sample of recent course offerings at the University of
Colorado
relating to neural and statistical computation. Consult department home
pages for more information about course topics.
- Artificial intelligence
- Neural networks
- Machine learning
- Reinforcement learning
- Advanced connectionist modeling
- Statistical pattern recognition
- Time series analysis and prediction
- Natural language processing
- Knowledge-based systems
- Robot control
- Speech processing
- Issues and methodologies in cognitive science
- Introduction to cognitive simulation
- Psychology of thinking and problem solving
- Judgement and decision making
- The scientific investigation of consciousness
- Language acquisition
- Brains, minds, and computers
- Neural systems
This page is maintained by Mike Mozer
(E-mail: mozer@colorado.edu)