A Call for Papers
to be included in a Special Session of the
Second International Conference on Computational Intelligence & Neuroscience.
The conference is part of the Third International Joint Conference on Information Sciences March 1 - 5, 1997, at Research Triangle Park, North Carolina,
USA
|
Special Session Title: Bridges between computational
and biological
intelligence Session Chair, Paul S. Prueitt, paul@htech.com,
fax: 301-306-8201
We will consider for presentation papers at the cutting edge of
experimental and theoretical investigations into the nature of machine and
biological intelligence. Summaries will be peer reviewed and are due by December
6th, 1996. Please submit summaries by fax or e-mail. Papers accepted for
publication will be given a full 40 minutes for presentation.
Peer review will be completed by January 1st, 1997.
Full papers are to be submitted by March 4.
Discussion paper:
Human reasoning deals with open quasi-formal structures,
and handles
nonstationary problems. Deep problems in machine
intelligence are solved by
the mechanisms underlying biological intelligence.
For example, the use of
conserved principles by living systems leads to methods
for data compression
into small knowledge sources that demonstrate a compression
inverse. Neural
networks, genetic algorithms, new logics (like quasi
axiomatic theory, fuzzy
logic) and set theory (like rough sets and fuzzy
set theory) reflect these
conserved principles, and suggest that computational
processes should be able
to solve the distributed data fusion problem in such
a way as to allow
information, imprisoned in large databases, to be
accessed as situationally
specific responses to user questions.
Evidence for the possibility of competent machine
based situational
analysis, based on specific knowledge, is found by
considering the
simplification of chemical science during and after
the discovery of the
atomic periodic table. By analogy, knowledge based
systems are generically
designed to discover the periodic table of any natural
complex system in a
stable mode. Once this table is discovered, the
situational analysis is much
easier than before. This discovery might be conducted
in such a way that
deterministic (fully constrained) logic and non-deterministic
(plausible)
reasoning are maintained separately.
Yet, situations are not just an aggregation of basic
elements into
ensembles. For situational analysis to handle non-stationary
problems, a
measurement of the world at a specific time and place
must be made. This
suggests that the measurement problem can be studied
as a cycle with assembly
stability disassembly phases where pattern extraction,
memory and behavioral
responses each involve the phenomena of cross scale
entanglement.
What is the minimal set of issues that are required
to support machine
intelligence? Can we produce computational examples
that illustrate these
issues? The mechanisms supporting biological intelligence
must be built from
the mechanisms of the world. For example, the architectural
issue of cross
scale entanglement might be mapped to the connectivity
of cortical columns.
As a generic mechanism, geometric filters and/or
topological transformations
might create complimentarity effects as well as space
filling convolution of
information - consistent with notions of holonomic
processes and synergetics.
Specific Issues of interest:
Papers on how non-invariance is fully resolved in
a Chaotic/Stability
boundary as mediated manifold emergence characterized
by a reduction from a
very high dimensional space to a much lower but still
high dimensional space.
(Primary references, research of W. Freeman)
Papers on mechanisms for encoding specific finite
classes of natural kind and
specific qualitative structure-activity relationships
(QSAR). (Primary
references, research of Victor Finn)
Papers on categorical invariance used to capture
the conserved
structure-activity relationships to be found as "period
tables" hypothesized
to be specific to any complex system under investigation.
This is the
Process Compartment Hypothesis (PCH). (Primary references,
research of Stuart
Kauffman, Paul Prueitt).
Papers on multi-level QAT, with a generic theory
of cross scale entanglement
to address the unstable transition of compartments
between stable modes, a
transition that is necessarily open to environmental
interaction
(entanglement). (Primary references, research of
Peter Kugler, G. V. Tzvetkova, D. Pospelov.)
Papers on the experimental evidence that the micro
process in the dendritic
fields of the various brain regions give rise to
representational spaces,
with electro-chemical circuit phase coherence arising
from quantum level
interaction with synapse function and neuronal discharges
at the axon
hillock. (Primary reference, research of Karl Pribram)
Papers on quantum level "non-measured states"
(the technical term, in quantum
mechanics (QM), for an "unmeasured state"
is a "beable") and measured states
(observables, in QM terms), and generalized quantum
cross scale entanglement
leading to macro-phenomenon produced through symmetry
induction in
perceptional systems. (References, Stuart Hameroff,
Karl Pribram, Peter Kugler,
Lere O'Shakunle.)
Papers on the emergence of order through a balance
of (deterministic) chaotic
dynamics, the aggregation of basic elements (features
of a substrate), and
the presence of a real time pragmatic axis (regarding
the uniqueness of
things). (References, Stuart Kauffman, C. S. Peirce,
et al)
Papers that evaluate possible hardware evolution
to conform to the properties
of distributed processing. Hardware systems; respectively,
neurocomputer,
optical computer and computers based on a asynchronous
stochastic processing
are needed to support a new paradigm combining neural
network architectures
with optical hardware, stochastic processes and experimental
evidence.
(References, N Farhat, T. Kohenon, B. Gruenwald,
et al.)
Papers on non Von Neumann architectures producing
small compact knowledge
sources that can be combined with QAT and knowledge
engineering to conduct
domain specific investigations. (Reference D. Pospelov,
G. Osiplov, V. Finn,
et al)
Also: Papers on language-neutral concept spaces.
Papers on integrated
hardware/software architectures for knowledge processors.
Papers on
associative network semantics that reflect relationships
among concepts or
ideas as expressed in a specific situation. Papers
on the methods for
validation of responses and knowledge representations.