The intellect of humans differs from that of machines. Our ability to
data-mine is based on getting input actively. We shall present an active
system exhibiting characteristics reminiscent to those of mid visual
recognition. We shall then identify brain signals related to psychological
properties such as the level of alertness using data-mining techniques.
The second part of our talk will focus on cell dynamics. We shall present a
dynamic gene network explicating the circadian biological clock in the
framework of system biology. We then analyze the processing ability of some
genetic networks as is quoted in Physics Today. We close our talk by
presenting a particular, though very basic, gene-like code in a cell, and
demonstrating how via basic operations of gene expression, a stem cell
develops into a functional organism. A magnificent property of the
resulting organism is its unprecedented ability to self-repair. This
technique of self-construction is applicable to nanotechnology self assembly.
------------------------------------------------------------------------------------------------------
Professor Siegelmann is one of the leading persons in the field of
nature-based computation and biological information processing systems.
Siegelmann incorporates the study of biological information systems on one
hand and techniques of machine learning and mathematics on the other hand
to both better model biology and to use these ideas in the built of model
artificial systems.
Siegelmann was the PI on the "Chaos Computer" BSF grant and collaborated
with Jim Yorke. She has contributed heavily to the field of neural networks
and her theorems are described as the basic theoretical foundations in this
field. Her work on "Computation Beyond the Turing Limit" appeared in the
journal Science with her sole authorship, and followed by the book: H.T.
Siegelmann, Neural Networks and Analog Computation: Beyond the Turing
Limit, Birkhauser, Boston, 1998, that is used in graduate seminars in
various leading universities.
Her contributions to the field of analog computation are summarized in a
series of papers. In PRL she lays out a novel analysis techniques enabling
to compare time efficiency between analog and discrete algorithms. In
PhysicaD she reveals the connection between the question of P vs. NP in the
analog domain to the existence "fully" chaotic attractors in nature. Her
recent 2003 paper in the Journal of Complexity describes a practical
applications of Probabilistic analysis of a differential equation for
linear programming.
Siegelmann has steered the theoretical foundations of bio-computation into
algorithms for clustering and data-mining as well as for memory
organization, based on findings of the neural system. She has applied these
algorithms in various systems mainly in the fields of medical informatics
and bio-informatics, but also in security and surveillance applications.
Siegelmann has recently published a couple of results on Gene Networks. Her
recent work on this topic was described in Physics Today in January 2004 as
a breakthrough, and establish her as a leader in bio-computation.
Professor Siegelmann is the director of the laboratory of bio-computation
in the Computer Science department in Umass Amherst. She is also a core
member of the cognitive sciences program. In addition she takes a
university level role as the Chair of BIGIALS (BioInformatics, Genomics,
and interdiscip. approaches to the Life Sciences). She is invited to give
talks in numerous conferences and institutions on unconventional models of
computation and in particular on the biologically motivated computers. She
is on the editorial board of both neural networks and chaos journals.
------------------------------------------------------------------------------------------------------
Prof. Hava Siegelmann E-mail: hava@cs.umass.edu
Department of Computer Science Phone: 413-57 S HAVA
UMass Amherst Secretary phone: 413-545-2744
Amherst MA 01003 www.cs.umass.edu