home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!gatech!darwin.sura.net!jvnc.net!rutgers!ub!acsu.buffalo.edu!ambati
- From: ambati@acsu.buffalo.edu (Balamural K. Ambati)
- Newsgroups: comp.ai.neural-nets
- Subject: A couple of questions
- Message-ID: <BuF6vE.LMs@acsu.buffalo.edu>
- Date: 11 Sep 92 15:26:01 GMT
- Sender: nntp@acsu.buffalo.edu
- Organization: UB
- Lines: 18
- Nntp-Posting-Host: lictor.acsu.buffalo.edu
-
- A couple of questions. Are neurons (biological) analog devices or
- ultimately finite machines at some quantal level? Specifically, if they
- were analog machines then is not the amount of information they can
- store, process, etc. essentially infinite? Is there any way to describe
- the computational ability/capacity of a group of N neurons (I mean can
- it be bounded, say less than e^(kN), for some constant k)?
-
- There seems to be a lot of work being done with chaos and neural networks.
- Is chaos theory generally accepted by neuroscientists as being able to
- possibly describe network behavior? I have read that in epileptic conditions,
- the Hausdorff (and also correlation) dimension of the behavior of the
- neuronal population is quite small compared to instances in which not all
- the neurons are firing. Does this mean, as might be intuitive, that during
- epilepsy, the neurons are doing very little computation?
-
- Your thoughts are appreciated.
- Balamurali K. Ambati
- ambati@ubunix.buffalo.edu
-