Information-driven nonlinear nanoengine hierarchies for biomimetic 'evolware'
INT -
International Nanobiological Testbed Ltd., R&D Dept.
Both in information electronics and
the aerospace field there is an increasing tendency to research and development
efforts for biomimetic smart and even evolvable hardware (the 'evolware')
for a number of applications, ranging from the achieving of an increased power
in information processing, possibly beyond the Turing limit, to the attaining
of more efficient structural health monitoring and of autonomous automata for
space exploration. This paper argues that the fundamentals on which the current
attempts are based would fail the biomimicry purposes, metaphorical
though they may be, as they miss the basic features of evolutionary system
dynamics that living systems are likely to embody according to some nanoscale
bioscience views. The current attempts, based just on formal logic, i.e. on a purely syntactic way of
information processing, would just give automata much more complicated than
those designed and built toward the end of the 18th century but not more 'evolutionary'
than them. The basic logical features of a system featuring biomimetic self-organizing
evolutionary dynamics are set forth. A discussion is given of the mathematical
problems met with by biological systems in case they would be supposed to
reason and/or evolve through a syntactic combinatorial logic, that would lead
to non-computability in practice from a cybernetic combinatorial
explosion; as a concrete example of the fact that the human brain with the aid
of tools created by its own ingenuity (computers and measuring instruments)
goes around such non-polynomial type of non-computability, the case of the work of flight controllers in airports
with heavy traffic is analyzed and shown to be genuine 'biological
calculation', carried out well beyond the basic limit to information processing
rate set forth by H.J. Bremermann for any system made up of atoms and
electrons. It is shown that the physics of a biomimetic evolutionary automaton
can be understood and properly implemented by describing its information
processing in the automaton's phase space, distinguishing interactions into
purely parametric (just strength-dependent) and coded (space arrangement and/or
frequency-dependent), and that the information processing evolution in the
phase space is characterized not by syntax in itself, but by the closed
chain
syntax Û
semantics Û
pragmatics
that physically corresponds to
production of entropy which is rejected into the automaton's environment. Such semantic
automaton, which can 'simulate' its environment through extralogical
compression of information and acts on the environment accordingly, just like
living systems do in their evolution, is shown to overcome the formal
logic paradoxes of self-reference corresponding to Gödel incompleteness
theorem, to Turing's 'halting problem' for recursive machines, and to Richard's
paradox for von Neumann's self-reproducing automata. Statistical mechanics
of this open system, where a convolution between the environmental time series
and the inside nonlinear dynamic ongoings occurs, shows that its physical
implementation must be based on nanostructures, working under far from
equilibrium conditions, and organized as hierarchies of nonlinear nanoengines
realizing a mixed Hamiltonian and dissipative chaotic dynamics
going from the nanoscale up to the macroscopic level. Microscopic physics
underlying the nanoscale level is shown to be involved in that dynamics and in
solving the hierarchy's intra- and inter-level communication problems, possibly
through an interplay based on quantum tunnelling and on the recently introduced
paradigm of 'quantum holography'
that overcomes the formal logic 'cybernetic crisis', caused by
exponential complexity, and the mentioned self-reference paradoxes by
introducing extralogical data-compression and hence semantics through
synchronous cooperativity by linear superposition throughout the levels, and by
nonlinear mapping through the nilpotent Heisenberg Lie group, from the smallest
size to the whole scale of the system. Together with said chaotic information
processing, quantum holography realizes physically the fundamental features of
biomimetic evolutionary dynamics: 1) coupling of the automaton with the
environment through extralogical compression of data; 2) no pre-programmed
algorithmic instructions for the automaton, but just preparation to learn; 3) no
net hardware - software distinction, but the genuinely biomimetic
physical-morphological solidarity of structure-function which is made
possible just on the nanoscale level, where mechanical and electronic
degrees of freedom can couple strongly into nonlinear information-driven
nanoengines and their hierarchies; 4) accordingly, geometric encoding of
information as patterns of energy, not bits, and lossless phase
gating in quantum holography; 5) robustness to noise and even the
capability of exploiting noise as a source of evolution.
© 1995-2008 Kazan State University