From August 1999 till November 2001 I worked at BBN Technologies,
Speech and Language
Processing department. I worked on text
understanding and information extraction technologies.
Before that I taught Computer Science at Northeastern University
for two years, pursuing a combined Math/CS graduate program at the same time.
My thesis is in Algebra & Symbolic Computation
and deals with algorithms on finite groups (permutation groups and matrix
groups over finite fields).
I like fixing software that doesn't work exactly right.
Being a Linux user, I can do it too :-) [I
pity Windoze users -- they mostly have to put up with whatever M$
thinks is best for them]. Here are some of my little
hacks for fellow Linux users.
I am looking for a new job:
My research interests:
- XML-based semantic representation of structured documents
and other objects on the Internet.
The web will eveolve to accomodate more semantics,
more structure in published documents. We need open and very
flexible standards to make this possible. Today's web is full
of knowledge, put there by its users empowered by hypertext and
simplicity of publishing; the problem is finding and cataloging
that knowledge, as well as supporting efficient collaboration
between publishers. Semantic markup will give the hyperlink
graph more structure,
while the ability to define and transfer complex objects
will offer new services and richer collaboration on the web.
- Symbolic computation, computational group theory.
My Ph.D. thesis was on
computations with the so-called black box groups
(when the information available about a finite group is minimal,
or the available representation is far from the natural one).
- Symmetry and search on algebraic and combinatorial objects.
I am interested in applying algebraic techniques to make search
that involves objects that have group structure faster.
- Internet security and network intrusion detection. I am
interested in using statistical machine learning methods
to detect potential attacks. While "cybercrime" is a hugely
overblown issue (partly out of lack of elementary tech knowledge
in the media, partly for political purposes by those pursuing their
own goals), and "cyberterrorism" luckily still belongs to the
realm of fiction, computer security is an important and
interesting research field, and cracker intrusions can be a
Some (hopefully) useful links:
Is the computer and 'net revolution unstoppable?
On the contrary, we are going to see a significant slowdown
in innovation, due to
software patents and
- Hollywood-financed legislation.
This [SSSCA, now CBDTPA, a bill by Fritz Hollings, "Senator from
Disney"] law would effectively announce the end of innovation in
digital technology: it would become illegal to create any hardware or
software that failed to incorporate controls chosen by the content
Andy Oram, O'Reilly editor
Links below this line remain from the time I was teaching at
the NU College of Computer Science.
My LISP students: find a Windows LISP interpreter (xlisp by
David Betz) here.
Here is a zipped "New Hacker's Dictionary"
(as a Windoze help file). Don't confuse hackers and crackers.
Read about the glorious early days when hackers organized hard drive
races and programmed their LISP machines to call the elevator.
BTW, this version (3.20) is quite out of date. The latest version (4.10) is
Some classes I used to teach at NU CCS:
- COM 1105 Computers and Applications
- COM 1100 Fundamentals of Computer Science
- COM 1101 Algortihms and Data Structures I
- COM 1201 Algortihms and Data Structures II