Machine Learning applied to Text, Information Retrieval and
Extraction. Since 1996 I have been working on
statistical approaches to text classification, clustering and
extraction. I was a member of the
CMU Text Learning Group and Tom Mitchell's World-Wide
Knowledge Base Project.
I was the leader of the project at JustResearch that created
Cora, a domain-specific
search engine over computer science research papers. It currently
contains over 50,000 postscript papers. You can read more about our
research on Cora in our IRJ journal paper or
a paper
presented at the AAAI'99
Spring Symposium. The Cora team also included Kamal Nigam, Kristie Seymore, Jason Rennie, Huan Chang and Jason Reed.
I am the author of rainbow, (and
its library, libbow), a LGPL'ed software
package for statistical text classification written in C.
I have been invited to give a tutorial at the Neural Information Processing Systems conference (NIPS*2002). The title is
"Information Extraction from the World Wide Web".
With Lillian
Lee, Tony
Jebara and Kamal
Nigam, I co-organized IJCAI'2001 workshop titled Text Learning: Beyond Supervision.
With Thorsten
Joachims, Mehran Sahami
and Lyle
Ungar, I co-organized a IJCAI-99 workshop on
Machine
Learning for Information Filtering.
With Rich Caurana, Virginia de Sa and Michael Kearns, I
co-organized a NIPS*98
workshop on "Integrating Supervised
and Unsupervised Learning".
With Mehran
Sahami, Mark Craven
and Thorsten
Joachims I co-organized a ICML/AAAI-98 workshop on Learning for Text Categorization.
Reinforcement Learning---especially with hidden state
and factored representations. My thesis
uses memory-based learning and a robust statistical test on reward in
order to learn a structured policy representation that makes
perceptual and memory distinctions only where needed for the task at
hand. It can also be understood as a method of Value Function
Approximation. The model learned is an order-n partially
observable Markov decision process. It handles noisy observation,
action and reward.
It is related to Ron, Singer and Tishby's Probabilistic
Suffix Trees, Leslie
Kaelbling's G-algorithm and Andrew Moore's
Parti-game. It is distinguished from similar-era work by Michael Littman, Craig
Boutilier and others in that it learns both a model and a policy,
and is quite practical with infinite-horizon tasks and large state and
observation spaces. Follow-on or comparison work has been done by
Anders Jonsson, Andy Barto,
Will Uther,
Leslie Pack Kaelbling,
Natalia Hernandez, and Sridhar Mahadevan.
The algorithm, called U-Tree, was demonstrated solving a
highway driving task using simulated eye-movements and deictic
representations. The simulated environment has about 21000 states,
2500 observations, noise and much hidden state. After about 2 1/2 hours
of simulated experience, U-Tree learns a task-specific model of
the environment that has only 143 states.
