From: Amara D. Angelica (amara@kurzweilai.net)
Date: Sun Mar 02 2008 - 20:30:50 MST
Interesting discussion. Some updates: According to S. Hagan et al [1],
"Tegmark considers in his paper two different scales at which quantum
computation might occur in the brain--one involving superpositions of
neurons firing and not firing (with calculated decoherence times of 10^-20
s), and another involving microtubules (calculated decoherence times of
10^-13 s). We agree that superpositions at the level of neural firing are
unlikely, and in fact play no role in the orch. OR or any other contemporary
quantum model. In the orch. OR approach, neural firings are entirely
classical, though they may be initiated by the outputs of microtubule
quantum processes in the neuronal interior. We therefore focus our attention
on Tegmark's assertions regarding decoherence times for
microtubule-associated quantum superpositions....Recalculation after
correcting for differences between the model on which Tegmark bases his
calculations and the orch. OR model (superposition separation, charge vs
dipole, dielectric constant) lengthens the decoherence time to 10^5 to 10^-4
s;; decoherence times on this order invalidate the assumptions of the
derivation and determine the approximation regime considered by Tegmark to
be inappropriate to the orch. OR superposition."
Hameroff expands on this in a chapter in a recent book, The Emerging Physics
of Consciousness [2]. As I have noted in Kurzweil Accelerating Intelligence
News [3], "A new experiment has come close to detecting quantum effects in a
macroscopic object. NSA physicists have measured the vibrations of a tiny
nanoelectromechanical arm [mass of 10^-16 kg] to probe the limits at which
quantum behavior breaks down and classical physics takes over..." [4]
Another interesting hypothesis for neural computations involves microtubule
tubulin dimers, which, according to my co-inventor Jack Tuszynski (prof. of
physics, U of Alberta and editor of The Emerging Physics of Consciousness
[5]), are thought to operate at 10^27 operations per second in an entire
brain (compared to 10^16 to 10^18 in other models). I'm in ongoing
discussions with Hameroff, Tuszynski (prof. of physics, U of Alberta),
Tegmark, and others for an article on this subject and I welcome all ideas.
[1] Quantum computation in brain microtubules: Decoherence and biological
feasibility. S. Hagan,. 1. S. R. Hameroff,. 2. and J. A. Tuszynski, Physical
Reviews E, 65: 061901, 2002.
[2] J. A. Tuszynski (Ed.), The Emerging Physics of Consciousness, Springer,
2006. From the introductory chapter by Tuszynski and Woolf: "Penrose and
Hameroff (1996) have put forth a highly disputed model of consciousness
based on quantum computation in microtubules within the brain's neurons.
This and other quantum models elucidate a number of enigmatic features of
consciousness; however, a few hurdles remain in establishing their
likelihood. Some of these difficulties are identified when designing
prototype quantum computers. One such obstacle is that quantum computers
will require a high degree of isolation from decoherence effects of the
local environment, or alternatively some kind of fault-tolerant architecture
that permits delicate quantum computing in the presence of realistic levels
of decoherence (Knill, 2005). The brain operates at body temperature, its
mass comprises 60 percent water, and is electromagnetically, chemically and
mechanically noisy, all of which would seem to severely shorten the time
allowed for quantum computation. Long-lasting, large-scale quantum states
are deemed to be impossible in the brain because a single ion, photon, or
thermal vibration can cause decoherence and hence random reduction to
classical states. On the other hand, proponents of a quantum approach to
consciousness point to a number of physical mechanisms in the brain that may
lengthen the time of quantum coherence and provide necessary quantum
isolation. Firstly, microtubules may be able to perform quantum computations
at room temperature because basic maintenance of microtubules is energy
dependent resulting in energy being continuously pumped in and out. This
situation is analogous to that of lasers, which work accordingly to quantum
optical principles at room temperature (see Hagan et al., 2002; Mershin et
al., 2004a). Secondly, the water of hydration surrounding microtubules
appears to be in an ordered state, which decreases noise (Hagan et al.,
2002). Thirdly, topological error correction (in a manner similar to that of
the fault-tolerant architecture described above) may protect delicate
quantum states (Hagan et al., 2002)."
[3]
http://www.kurzweilai.net/news/frame.html?main=news_single.html?id%3D3157
[4] M. D. LaHaye et al, Science 2 April 2004 304: 74-77, DOI:
10.1126/science.1094419,
http://www.sciencemag.org/cgi/content/full/304/5667/74, accessed on the Web
3/2/08.
[5] From J. A. Tuszynski (Ed.), The Emerging Physics of Consciousness:
"There are on the order of 10^4 synapses per large neuron, which switch
their states at a rate of some 10^3 switches per second, so that we arrive
at a number of ~10^18 operations per second in the brain on average. While
this is a truly huge number, it may pale by comparison with the yield given
by the brain if neuronal microtubules were actively involved in
computational processes. Consider that at the cytoskeletal level there are
roughly 10^7 microtubule tubulin dimers in each neuron which can switch
their conformational states on the order of nanoseconds resulting in on the
order of 10^16 operations per second per neuron or 10^27 operations per
second in an entire brain instead of 10^18 operations per second estimated
for the coarse-grained approach where neurons are taken as the smallest
computational units. Moreover, if each tubulin dimer does function as a
qubit and not a classical bit processor, then the computational power
becomes almost unimaginably vast. It has been claimed that as few as 300
qubits have the same computational power as a hypothetical classical
computer comprised of as many processing units as there are particles in the
universe."
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:02 MDT