--- "John G. Rose" <[EMAIL PROTECTED]> wrote: > > From: Matt Mahoney [mailto:[EMAIL PROTECTED] > > > > That's true. The visual perception process is altered after the > > experiment to > > favor recognition of objects seen in the photos. A recall test doesn't > > measure this effect. I don't know of a good way to measure the quantity > > of > > information learned. > > > > When you learn something is it stored as electrical state or are molecules > created? Perhaps precise measurements of particular chemicals in certain > regions could correlate to data differential. A problem though is that the > data may be spread over a wide region making it difficult to measure. And > you'd have to be able to measure chemicals in tissue structure though > software could process out the non-applicable.
I was thinking more in terms of theoretical information capacity independent of implementation. Landauer's experiments measured memory using recall tests. This measure is more useful because it tells you how much memory is needed to reproduce the results in a machine. The brain may be doing it less efficiently but we don't need to implement it the same way. If a person can memorize 100 random bits, then you can be certain that any implementation will need at least 100 bits of memory. > But a curious number in addition to average long term memory storage is > MIPS. How many actual bit flips are occurring? This is where you have to be > precise as even trace chemicals, light, temperature, effect this number. > Though just a raw number won't tell you that much compared to say > spatiotemporal MIPS density graphs. A molecular level simulation of the brain is sure to give a number much higher than the most efficient implementation. A more interesting question is how much memory and how many MIPS are needed to simulate the brain using the most efficient possible model? -- Matt Mahoney, [EMAIL PROTECTED] ------------------------------------------- singularity Archives: http://www.listbox.com/member/archive/11983/=now RSS Feed: http://www.listbox.com/member/archive/rss/11983/ Modify Your Subscription: http://www.listbox.com/member/?member_id=4007604&id_secret=96140713-a54b2b Powered by Listbox: http://www.listbox.com
