home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!gatech!paladin.american.edu!auvm!BEN.DCIEM.DND.CA!MMT
- Return-Path: <@VMD.CSO.UIUC.EDU:mmt@ben.dciem.dnd.ca>
- Message-ID: <9301211851.AA22536@chroma.dciem.dnd.ca>
- Newsgroups: bit.listserv.csg-l
- Date: Thu, 21 Jan 1993 13:51:25 EST
- Sender: "Control Systems Group Network (CSGnet)" <CSG-L@UIUCVMD.BITNET>
- From: mmt@BEN.DCIEM.DND.CA
- Subject: Re: Multiple directions; arm models
- Lines: 56
-
- [Martin Taylor 930121 13:40]
- (Bill Powers 930121.0830)
-
- >I can see perceptions being made of individual contributions from
- >many individual direction-detectors aimed in different
- >directions, but when it comes to output the available directions
- >are limited to the degrees of freedom of the arm.
-
- Of course. But which basis functions should be used to describe those
- degrees of freedom need not be fixed over time. That's the basis of the
- attention-fluctuation argument, after all. We used to be well agreed on
- that. Even if the output directions were fixed over time, it is highly
- unlikely that they would be fixed to a particular x-y direction in a plane
- at some arbitrary orientation with respect to the head and body.
-
- >Exactly orthogonal control is not necessary; if there is a slight
- >deviation from a 90-degree angle between the axes, motion in one
- >axis simply disturbs the other axis a little, and the control
- >system on the other axis cancels the disturbance by normal means.
-
- Naturally. No problem.
-
- >>So, to enhance the model, I would put more ECSs, which look at
- >>error in different directions, rather than impute to the human
- >>an ability to rotate the space.
- >
- >You'd better work out an example; this just looks like a great
- >opportunity for conflict to me.
-
- Well, it's the kind of conflict you have previously called "muscle tone."
- Not entirely a bad thing. In a well distributed network, there's conflict
- all over the place. We know that, and (in my view) it prevents the
- hierarchy from getting frozen, keeping at least a low level of reorganization
- going. It's the equivalent of annealing in conventional neural networks.
- Helps the control hierarchy from getting stuck in local minima.
-
- >>If one does make a model with more ECSs, there are more degrees
- >>of freedom with which to fit the data, which makes it more
- >>difficult to determine whether the new model is an improvement
- >>over the old.
- >
- >Slow down. First demonstrate that such a model would work at all.
- >Then worry about degrees of freedom.
-
- I think Rick is doing that. If he isn't, I will, using his stack, the
- revised version of which he has kindly sent me (but I haven't got it
- onto the Mac yet).
-
- Martin
-
- PS. Re Durango: would it be helpful to propose a voluntary fund for the
- purpose of subsidizing students to attend? I assume that the fee you set
- is calculated to allow for the three subsidies provided that you get the
- anticipated number of attendees. If you had an additional fund you wouldn't
- need to worry about how many people come. I'd be happy to contribute, and
- I would be surprised if you didn't find a few more.
-