Around 2001-2002, I worked for Quantum Leap Innovations,
a company that specializes in the development of intelligent software. The greater portion of my
work was in the implementation and evaluation of algorithms for constrained optimization problems
with discrete and/or continuous parameters.
Research Interests
Although I have a variety of interests within computer science, my main focus is towards the
disciplines of artificial intelligence (AI) and artificial life (ALife). I am especially interested
in the computational issues that arise when an entity attempts to deal with
uncertainty, misleading information and
rapidly-changing surroundings. These are difficulties that humans routinely face and overcome in
the most ordinary day-to-day situations, yet they pose a problem for present-day AI technologies.
I hypothesize that endowing machines with the capacity to overcome this problem will require an
in-depth understanding of various "non-traditional" computational models, which may hold properties
such as nondeterminism and massive parallelism. Many of these models take inspiration from phenomena
found in Nature, and include paradigms such as evolutionary computation, neural modeling and swarm
intelligence.
My dissertation research, which was conducted under the advisement of
Prof. James A. Reggia,
dealt with the application of swarm intelligence techniques to the
problem of self-assembly. Here, the agents are embodied as
different-sized blocks that must use local information to guide
themselves through a constrained, continuous space, such that they eventually
form some desired structure. I was able to design agent-level
control rules for achieving the self-assembly of 3D structures such as
the one illustrated in this sample video.
Further, I developed methods for generating such rules automatically,
given a desired structure. This work yielded a better understanding
of the highly complex relationship between individual, agent-level
behaviors and emergent, system-level properties.
J. Reggia and A. Grushin, Population Lateralization Arises in Simulated
Evolution of Non-Interacting Neural Networks (commentary),
Behavioral and Brain
Sciences28, pp. 609-611 (2005).
E. Nowak, A. Grushin, A. Barnum and M. Weissman, Density-Noise Power
Fluctuations in Vibrated Granular Media,
Physical Review E - Rapid
Communications63, 020301 (2002).
Apart from computer science, I have a variety of other interests,
which include (among other things) music, literature, poetry, film,
food and travel.
Random Thoughts
Sometimes, science shows us not only what we can do, but also what we cannot do, and I often find this
to be particularly fascinating. (An especially striking example, which will be familiar to
most computer scientists, is Gödel's Incompleteness Theorem).
Although I study artificial intelligence, I believe that intelligence
and consciousness are not the same thing, and that consciousness has
neither a physical nor a computational explanation...
