From newshub.ccs.yorku.ca!torn!cs.utexas.edu!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!destroyer!uunet!trwacs!erwin Wed Aug 12 16:52:10 EDT 1992
Article 6548 of comp.ai.philosophy:
Path: newshub.ccs.yorku.ca!torn!cs.utexas.edu!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!destroyer!uunet!trwacs!erwin
>From: erwin@trwacs.fp.trw.com (Harry Erwin)
Newsgroups: comp.ai.philosophy
Subject: Re: Memory and store/retrieve.
Message-ID: <684@trwacs.fp.trw.com>
Date: 3 Aug 92 12:41:40 GMT
References: <1992Jul30.152320.2247@puma.ATL.GE.COM> <1992Jul31.160209.26718@mp.cs.niu.edu> <BILL.92Jul31195028@ca3.nsma.arizona.edu> <1992Aug1.132812.12457@mp.cs.niu.edu>
Organization: TRW Systems Division, Fairfax VA
Lines: 25

>"if we did not have language, what would episodic memory be like?"
The most vivid of episodic memories appear to be keyed by emotional and
sensory states, particularly the sense of smell/taste. I think that
primitively, episodic memory is the association of a collection of
sensory states with an emotional state, with the emotional state serving
to calibrate a dynamic programming threshold associated with the one-sided
game with information collection.

One-sided games with information collection have threshold strategies--if
the current estimate of the probability of success exceeds that threshold,
the player should continue; otherwise he should abandon the game. The
current estimate is not expressed in a quantitative measure, but rather in
an emotional state. (Think of fight or flee reactions.)

Even the two-sided game will have a threshold strategy in most biological
contexts. Although the preferred strategy should evolve chaotically,
chaotic dynamics are only seen when behavior is labial. If the behavior is
either genetically encoded or calibration has non-trivial costs, it is
cheaper to use a threshold strategy and accept the reduced fitness.

Cheers,
-- 
Harry Erwin
Internet: erwin@trwacs.fp.trw.com



