dorettas@iddic.UUCP (Doretta Schrock) (10/22/85)
[pseudo-bug eater] I'm always excited when a new posting appears in net.cog-eng. Will it be about some new connectionist theory? Maybe some theoretical discussion of brain vs. computer? Or maybe a comment on a new model of perception, attention, cognition, or memory? but NOOOOOOOOOOOOOOOOOOOOOOOOOOOO... Recently this group has been degenerating into a clone of net.micro.mac, which I (and many others, judging by the volume) read. PLEASE keep your discussions of DRI vs. Apple (and etcetera) OVER THERE (or preferably, in your head)! There is so much going on in AI, cognitive engineering, neurology, and cognitive science (to name but a few), that we should be able to generate some interesting and useful dialog without stooping to discuss the same things over and over and over and over and over again. For example: Has anyone else read _Godel, Escher, Bach_? Or Pylyshyn's _Computation and Cognition_? Or _The Psychology of Human-Computer Interaction_? What did you think? Are the emerging connectionist theories going to revise our view of AI and BI (biological intelligence)? Should we all junk our present computers and languages for big multiprocessors? Will examining the way the Aplysia works really help us understand cognition (I read that you folks at ATT were doing something along these lines...or was that a wildly unfounded rumor?)? What is the best way to get a handle on what's going on in cog.eng. today? What should you study if you are in school, or what should you read if you are out? Will those of us without PhD's have a chance at any juicy industrial/ R&D jobs (no snickering from the grad students)? I could go on, but that ought to give some of you with fertile minds someplace to start. Really, this wasn't meant so much as a flame as it was a plea for more quality and less mediocrity in what's taking up space in this newsgroup. If you didn't "n" already, I thank you for your attention. Please send all flames to /dev/null or the round bit bucket, whichever is closer. Looking forward to some truly dazzling dialog, Mike Sellers <--note the difference in name from above
crs@lanl.ARPA (10/23/85)
> . > . > . > Recently this group has been degenerating into a clone of net.micro.mac, > which I (and many others, judging by the volume) read. PLEASE keep your > discussions of DRI vs. Apple (and etcetera) OVER THERE (or preferably, > in your head)! There is so much going on in AI, cognitive engineering, > neurology, and cognitive science (to name but a few), that we should be > able to generate some interesting and useful dialog without stooping to > discuss the same things over and over and over and over and over again. > . > . > . > Mike Sellers <--note the difference in name from above Yes! How about it folks? -- All opinions are mine alone... Charlie Sorsby ...!{cmcl2,ihnp4,...}!lanl!crs crs@lanl.arpa
doc@cxsea.UUCP (Documentation ) (10/25/85)
> [pseudo-bug eater] > > I'm always excited when a new posting appears in net.cog-eng. > Will it be about some new connectionist theory? Maybe some > theoretical discussion of brain vs. computer? Or maybe a comment > on a new model of perception, attention, cognition, or memory? > > but NOOOOOOOOOOOOOOOOOOOOOOOOOOOO... > > > I could go on, but that ought to give some of you with fertile minds someplace > to start. Really, this wasn't meant so much as a flame as it was a plea for > more quality and less mediocrity in what's taking up space in this newsgroup. Well, gee, Mike. I'm neither a cog, or an eng, but I'll try this: I've been learning about dyslexia. I know a few people who are dyslexic, and I've been attempting to devise some exercises for them on a Macintosh using MacDraw/Paint, as a means of mapping how they process and respond to visual cues. Has anyone else explored this? Any observations to share?
gilbert@hwcs.UUCP (Gilbert Cockton) (10/29/85)
In article <2246@iddic.UUCP> dorettas@iddic.UUCP (Doretta Schrock) writes: > >I'm always excited when a new posting appears in net.cog-eng. >Will it be about some new connectionist theory? Maybe some >theoretical discussion of brain vs. computer? Or maybe a comment >on a new model of perception, attention, cognition, or memory? > >Recently this group has been degenerating into a clone of net.micro.mac, >which I (and many others, judging by the volume) read. PLEASE keep your >discussions of DRI vs. Apple (and etcetera) OVER THERE (or preferably, >in your head)! This group does seem to have been colonised by the pragamatic school of Computer-Human Interaction. Personally I do find some of the ideas on interaction techniques interesting and I would certainly miss them. The problem seems to me to the group's name `Cognitive Engineering'. As far as this title is concerned, it is NOT a Cognitive Science group. As I see it the net.ai already carries a lot of cognitive science material, so perhaps you're looking in the wrong place when you read this group. Perhaps a new group called net.chi is needed that will cater for both theoreticians and practitioners in interactive systems, leaving the purists in cog-eng to discuss the length and shape of the brain along with its optimal processing speed in tachistoscopic experiments. >Has anyone else read _Godel, Escher, Bach_? Or Pylyshyn's _Computation and >Cognition_? Or _The Psychology of Human-Computer Interaction_? What did >you think? Yes, no, yes. Interesting, can't say, disappointing ( I would never recommend this as early reading for anyone new to hci, it is far too idiosyncratic and academically sloppy, I'd rather have 4,000 DOD CHI guidelines !)
eugene@ames.UUCP (Eugene Miya) (10/30/85)
[I know I'm going to catch hell from some of the psychologists who work here.]
Didn't theories of connectionism die out in the late 1950s and early 1960s?
Are some of the AI people of today reinventing neural nets? Didn't Feldman,
Minsky, Norman, and numerous others do all this and hash back and forth
between each other?
From the Rock of Ages Home for Retired Hackers:
--eugene miya
NASA Ames Research Center
{hplabs,ihnp4,dual,hao,decwrl,allegra}!ames!aurora!eugene
emiya@ames-vmsbhalff@utah-cs.UUCP (Henry M. Halff) (11/01/85)
> From eugene@ames.UUCP (Eugene Miya) Tue Oct 29 22:17:38 1985 > Subject: Re: cognitive-engineering, connectionism, etc. ...really flame-ola > Organization: NASA-Ames Research Center, Mtn. View, CA > > [I know I'm going to catch hell from some of the psychologists who work here.] > > Didn't theories of connectionism die out in the late 1950s and early 1960s? > Are some of the AI people of today reinventing neural nets? Didn't Feldman, > Minsky, Norman, and numerous others do all this and hash back and forth > between each other? > > >From the Rock of Ages Home for Retired Hackers: > --eugene miya > NASA Ames Research Center > {hplabs,ihnp4,dual,hao,decwrl,allegra}!ames!aurora!eugene > emiya@ames-vmsb > As I understand it hterw was a lot of early work on connectionist networks of a form called perceptrons, which were essentially one-dimensional teachable pattern recognizers. The work was brougth to an abrupt halt when Minsky and Papert (see their book, Perceptrons), proved that these devices could not learn patterns that exhibited certain simple dependencies. More recent research is looking at more complicated network architectures. Minsky and Papert's objections can be ovecome with devices that have several layers of units, but there ares till people who find it convenient to beleive that Minsky and Papert put the whole issue to bed. -- Henry M. Halff Halff Resources, Inc. halff@utah-cs.ARPA 4918 33rd Road, N., Arlington, VA 22207