home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!paladin.american.edu!darwin.sura.net!zaphod.mps.ohio-state.edu!magnus.acs.ohio-state.edu!sample.eng.ohio-state.edu!purdue!mentor.cc.purdue.edu!noose.ecn.purdue.edu!schmied.ecn.purdue.edu!surana
- From: surana@schmied.ecn.purdue.edu (Pankaj K Surana)
- Newsgroups: comp.ai.neural-nets
- Subject: GA+NN=AI
- Keywords: genetic algorithms, neural nets
- Message-ID: <1992Sep15.195413.14226@noose.ecn.purdue.edu>
- Date: 15 Sep 92 19:54:13 GMT
- Sender: surana@en.ecn.purdue.edu
- Organization: Purdue University Engineering Computer Network
- Lines: 23
-
- Here's a question: Is there any literature that discusses
- the possible use of genetic algorithms in conjunction with
- neural networks.
-
- I have been thinking that, perhaps, a genetic algorithm that
- manipulates the formation of neural nets could, through
- evolutionary changes, discover a more appropriate neural
- model for cognitive research. The "neuron" itself would remain
- unchanged, but the number of and structure of interconnections
- between neurons would be changed according to some algorithm.
-
- As another method, would it be possible for a neural structure
- to be created and implemented "on the fly", that is, as a stream
- of data is sent through it? Could the neurons align themselves
- appropriately to handle different types of data, with different
- resultant effects?
-
- Just a few questions thrown out to the group. Any thoughts should
- be e-mailed to surana@en.ecn.purdue.edu
-
- Thanks,
-
- Pinku Surana
-