From: Eliezer S. Yudkowsky (sentience@pobox.com)
Date: Sun Jan 30 2005 - 08:29:19 MST
Phil Goetz wrote:
> The problem is harder than is suggested by the 200Hz
> figure. That is a spike-firing frequency.
> In many cases, we believe info is encoded in the
> firing rate,
I have previously disputed this assumption, not least because of its
consequence. (Ten steps is just ridiculous...) Neurologists can easily
count spikes. Decoding information is much harder. Thus, I suspect that
many complex encodings are reported in journals as "firing rates" because
that's what neurologists know how to measure and they don't know what's
being computed, let alone how to decode it. For some brain areas, our
knowledge of neurobiology may be equivalent to measuring the "firing rates"
of transistors in a computer with no idea of what the transistors encode.
In other areas, especially vision, we have some idea of what is being
computed and so we can do better than that.
> so the number of sequential neuron
> firings used in reacting to a stimuli is probably an
> order of magnitude smaller. I think neurologists call
> this the "ten-step problem", meaning that the brain
> often needs to compute a response with only time
> enough for 10 sequential operations. I think Newell's
> book Unified Theories of Cognition says more about
> this.
-- Eliezer S. Yudkowsky http://intelligence.org/ Research Fellow, Singularity Institute for Artificial Intelligence
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:50 MDT