| |
|
|
The Artificial Intelligence group focuses on
intelligent computing. Our research objective is to make the computer
systems more intelligent, with the ability to hear and understand human
speech and languages, to learn from examples, to extract and recognize important
features from complex data, to represent and reason about knowledge in
various new and advanced programming languages, and to plan courses of
actions. The enabling technology areas that we study are machine
learning and data mining, speech recognition, logic programming and
Bayesian networks, natural language processing and information
retrieval, and computer music and audio engineering.
We continue to
disseminate our research results in high-profile international
conferences and journals, and apply our research to application areas
such as programming language design, Chinese medicine, biometrics,
speech and handwriting recognition and computer security, computer music
as well as data mining and its applications on the Web and in
bioinformatics.
Focused research areas
and projects
Seminar
Schedule
 Courses
UG
Courses:
COMP
221: Fundamentals of Artificial Intelligence
COMP 327: Introduction to Pattern Recognition
COMP 300Y: Introduction to Machine Learning
COMP 342: Introduction to Computer Music
PG
Courses:
COMP 521: Artificial
Intelligence
COMP 522: Machine Learning
COMP 526: Natural Language
Processing
COMP 527: Pattern Recognition
COMP 537: Knowledge Discovery
in Databases
COMP 538: Reasoning and
Decision under Uncertainty
COMP 621: Advanced Topics
in Artificial Intelligence
Focused
Research Areas
l
In knowledge representation,
we are working on Bayesian networks, answer-set logic programming, and
case-based reasoning
l
In machine learning and data
mining, we are working on Web mining to uncover Web users’
preferences and access patterns, clustering and analysis of sequential
data, semi-supervised learning, embedding of high-dimensional data,
transductive learning for IDS alarm processing, sensor fusion for
intrusion detection.
l
In planning, we are
studying how to learn to make better plans from past plans, with
application to customer relationship management and robotic
programming.
l
In speech and natural language
understanding, we apply statistical methods to enable more
effective techniques to understand human speech and translate between
different natural languages
l
In computer music and audio
engineering, we are working on Matching Synthesis of Western
and Chinese Musical Instruments Musical Instrument Recognition.
Research
Projects
|
HORNER, Andrew |
-
Wavetable Matching
with Modified Spectral Snapshots, Sept 2002 - Aug 2005
-
Wavetable Matching
Using Combinatorial Basis Spectra Selection, Sept 2001 - Aug
2004
-
Low Peak Amplitudes
for Wavetable Synthesis, Aug 1999 - July 2001
|
|
KWOK, James |
- Machine learning for
content-based and collaborative recommender systems
- Information geometry of
kernel methods with applications to data mining
- Support vector machine with
Bayesian learning for the automatic categorization of text and
hypertext documents
|
|
LIN, Fangzhen |
- Learning and its role in reasoning. Specifically,
I'm interested in provably correct learning methods to make
reasoning more efficient.
- Planning and learning. We are
using Planner-R to
experiment how learning can help planning. This is an ongoing RGC
CERG project.
- ASSAT - computing answer sets
by using SAT solvers.
|
|
MAK, Brian |
- Discriminative Training of Non-parametric
Auditory Filters for Automatic Speech Recognition, 2002-2004
- Asynchronous Multi-Band Continuous Speech
Recognition using HMM Composition, 2002-2003
|
|
WU, Dekai |
-
Statistical
Natural Language Processing
-
Machine
Translation.
|
|
YANG, Qiang |
-
Web
mining for Web user access-pattern prediction, with application to
Web caching, prefetching and e-learning 2001-- 2004
-
Data
Mining for case bases and plans, with application to customer
relationship management 2002—2004
|
|
YEUNG, Dit-Yan |
-
Semi-supervised learning
-
Embedding/manifold/spectral methods
-
Kernel methods
|
|
ZHANG,
Nevin |
-
Efficient Algorithms for Near-Discernible Partially Observable
Markov Decision Processes, September 2001 - August 2004
- Toward a Statistical Foundation for Traditional Chinese Medicine
(TCM) Diagnostics, January 2000 - present
|
Affiliated
Centers
and
Joint Projects
Cyberspace Center
Human Language Technology
Center (HLTC)
Virtual English Language Adviser Project (VELA)
|
