Daniel Lowd
|
E-mail: lowd at cs dot washington dot edu
Office: 430 Allen Center
Mailing Address:
University of Washington
Dept. of Computer Science and Engineering
Mailstop 352350
University of Washington
Seattle, WA, 98195-2350
|
I am a sixth-year graduate student in
Computer Science and Engineering at the University of Washington.
I expect to graduate in the summer of 2009. You
can find my C.V. here.
Research
My focus is on statistical relational learning, but I'm also
interested in recommender systems, spam filtering,
and machine learning in general. I am currently working with Pedro Domingos on
miscellaneous things that are somehow related to Markov logic networks.
Statistical Relational Learning
Statistical relational learning seeks to represent the complexity and
uncertainty present in most real-world problems by combining
first-order logic with probability. The main challenges are in
developing effective representations and effective algorithms. One of
my projects has been Recursive Random Fields (RRFs), a multi-layer
generalization of Markov logic networks that resolves a number of
inconsistencies in the Markov logic representation (Lowd and Domingos,
2007a
[pdf]
[ppt]
[ppt+audio]).
I have also worked on applying quadratic optimization
algorithms to Markov logic weight learning, resulting in more accurate
models in much less time than before (Lowd and Domingos, 2007b
[pdf]
[ppt]
[video]).
Learning for Efficient Inference
Inference in Bayesian networks and Markov networks is intractable in
general, but many special cases are tractable. Often, a tractable
subclass such as naive Bayes mixture models yields comparable accuracy
while offering exponentially faster inference (Lowd and Domingos,
2005
[pdf]
[ppt]
[appendix]).
Furthermore, by incorporating a preference for tractable
models into the learning algorithm, we can guarantee efficient
inference without restricting ourselves to any particular class (Lowd
and Domingos, 2008
[pdf]
[pdf+proofs]).
Adversarial Machine Learning
I spent the summer of 2004 at Microsoft Research working with
Chris Meek on
the problem of spam. We looked at a common technique spammers use to defeat
filters: adding "good words" to their emails. We developed techniques
for evaluating the robustness of spam filters, as well as a theoretical
framework for the general problem of learning to defeat a classifier
(Lowd and Meek, 2005ab
[pdf]
[pdf]).
Slides from a talk at Oregon State University
(7/14/2006).
Slides from a talk at the 2007 NIPS
Workshop on Machine Learning in Adversarial Environments for Computer
Security (12/8/2007).
Slides from a talk at the
University of Cagliari, Italy (7/3/2008).
Publications
Book chapters
- Just Add Weights: Markov Logic for the
Semantic Web.
Pedro Domingos, Daniel Lowd, Stanley Kok, Hoifung Poon, Matthew
Richardson, and Parag Singla.
In P. C. G. Costa, C. d'Amato, N. Fanizzi, K. B. Laskey, K. J. Laskey,
T. Lukasiewicz, M. Nickles, and M. Pool (eds.),
Uncertain Reasoning for the Semantic Web I, 2008.
New York: Springer.
- Markov Logic.
Pedro Domingos, Stanley Kok, Daniel Lowd, Hoifung Poon, Matthew
Richardson, and Parag Singla.
In L. De Raedt, P. Frasconi, K. Kersting and
S. Muggleton (eds.), Probabilistic Inductive Logic Programming (pp.
92-117), 2008. New York: Springer.
Refereed conference papers
- Using Salience to Segment Desktop Activity
into Projects.
Daniel Lowd and Nicholas Kushmerick. Proceedings of the Thirteenth
International Conference on Intelligent User Interfaces (IUI), 2009.
Sanibel Island, Florida: ACM Press.
- Learning Arithmetic Circuits.
Daniel Lowd and Pedro Domingos. Proceedings of the Twenty-Fourth
Conference on Uncertainty in Artificial Intelligence (UAI), 2008.
Helsinki, Finland: AUAI Press.
(Extended version with proofs)
(Poster)
- Efficient Weight Learning for Markov
Logic Networks.
Daniel Lowd and Pedro Domingos. Proceedings of the Eleventh
European Conference on Principles and Practices of Knowledge
Discovery in Databases (PKDD), 2007. Warsaw, Poland: Springer Verlag.
(Slides)
(Video)
[Updated PDF file fixes several formula errors.]
- Recursive Random Fields.
Daniel Lowd and Pedro Domingos. Proceedings of the Twentieth
International Joint Conference on Artificial Intelligence (IJCAI), 2007.
Hyderabad, India: IJCAI. (Slides)
(Slides+Audio)
- Naive Bayes Models for Probability Estimation.
Daniel Lowd and Pedro Domingos. Proceedings of the Twenty-Second
International Conference on Machine Learning (ICML), 2005. Bonn, Germany:
ACM Press. (Slides)
(Online appendix)
- Good Word Attacks on Statistical Spam Filters.
Daniel Lowd and Christopher Meek. Second Conference on Email
and Anti-Spam (CEAS), 2005. Palo Alto, CA. (Slides)
- Adversarial Learning.
Daniel Lowd and Christopher Meek. Proceedings of the Eleventh ACM SIGKDD
International Conference on Knowledge Discovery and Data Mining (KDD), 2005.
Chicago, IL: ACM Press. (Poster)
Workshop papers
- Recursive Random Fields (workshop version).
Daniel Lowd and Pedro Domingos. Proceedings of the ICML-2006 Workshop on
Open Problem in Statistical Relational Learning, 2006. Pittsburgh, PA: IMLS.
(Slides)
Technical reports
Miscellaneous
CSE Poetry
Here is a villanelle I wrote for architecture class. I had to
write it in order to get a one-week extension on the final project. Writing
was quite enjoyable... alas, there is no quals course in poetry!
Fortunately, one of the questions on the architecture final asked me to
answer a question of my own creation. I received
full credit on it, too.
The following quarter inspired this creation.
I was taking advanced
complexity at the time.
Music
World of Warcraft meets the Beach Boys: Ungoro
(MP3). Lyrics, vocals, and production by me.
I used to sing with the Seattle Men's
Chorus.
Other Interests
My wife, Mary Lowd, is a science
fiction writer. You can read one of her short stories
online for free. Her web page includes some excellent pictures of our
daughter
and our pets.
I don't plan on ever getting a tattoo, but if I did, it would have to
be something really cool. Like a praying mantis. Playing a guitar.
Eating a butterfly.