-- Messages from file: [SUMEX-AIM]<LENAT>MESSAGE.TXT;1
		 -- Sunday, January 7, 1979 13:09:17 --

   9   20 Nov  Rick at Rand-Unix     eply to doug's comments
   10  20 Nov  Klahr at Rand-Unix    Cognitive Economy Paper
   12  20 Nov  Rick at Rand-Unix     Re: (Response to message)
   150 26 Dec  Rick at Rand-Unix     joint paper


9 -- ************************
Mail from RAND-UNIX rcvd at 20-Nov-78 1218-PST
From: Rick at Rand-Unix
Date: 20 Nov 1978 at 1218-PST
Message-Id: <[Rand-Unix]20-Nov-78 12:18:32.rick>
To:     lenat@aim
cc:     klahr,rick
Subject:Reply to doug's comments

Doug:
	Am I to infer that you want to first author badly enough
that you were unwilling to send ME your comments.  Personally,
I'd be happy to have you be first author given the responsibility
it carries.  Now it's between you and Phil (so send your messages
to me only, and I'll forward them to him).

	My replies to your comments are indicated by **.

				-Rick



------- Forwarded Message

From: Klahr at Rand-Unix
Date: 20 Nov 1978 at 1033-PST
Message-Id: <[Rand-Unix]20-Nov-78 10:33:05.klahr>
To: rick
cc:
Subject: Doug's Comments

I didn't see your name in Doug's message, so I'm forwarding his comments.

------- Forwarded Message

Date: 18 Nov 1978 1848-PST
From: Doug Lenat <DBL at SU-AI>
Subject: Caching  
To:   klahr at RAND-UNIX    

Rick and Phil,

Thanks for the notes.  I like Rick's outline, and have some comments
on it and on Phil's remarks as well.


First, title and authors.  In some sense this is not important, but
pragmatically it can have a big impact on the way a paper is received.
How about just Cognitive Economy?  This enables the paper and its
**this title is fine

concepts to be referred to tersely, and will pique the curiosity of
many potential readers.  "Economics" sounds too presumptuous for the
material we have; the longer titles Phil suggests are more precise but
unwieldy.                 For author order, I would like to go first,
but if there is dissention we can draw lots or something.  The primary
author should have the brunt of the task of preparing the article,
in any case, including content, prose, and document preparation.
** i'm happy with this.  however there are some ideas from psychology that
** may not have.  I'm sending you two of our previous psych articles
** that bear on the thesis.  One claims that people reason in terms of
** words themselves, not primitives, because the higher-level code
** is more efficient.  The other shows that people develop direct associations
** between co-retrieved items and this determines the so-called
** "semantic" net in human memory.  Hence "dog" and "animal" are
** closer than "dog" and "mammal" (although dog < mammal < animal).

The thesis that Rick states is much too strongly worded for my taste;
co-authoring with Rick is always a pleasure because he's the only
person as brash as I -- and occasionally moreso!  But I really don't
want to claim that we have  the final set of three keys (abstrac.,
caching, expectation-simplified processing) to represenational economy.
In fact, can we think of more?
** how many more do i have to think of without proper remuneration??
** how about one dinner per key?
** anyway--if you want to influence lot's of people with this paper,
** i suggest we choose a better term for caching...it's too narrow a
** metaphor (since it include's compiling, recoding, storing... all
** in response to retrieval opportunities and requests).  How about
** "recording" or "banking" or "stockpiling" or "archiving" or ...

THESIS:  As often happens when I read Rick's stuff, I misunderstood it the
first time through, and as a result got an additional good idea out
** but did you eventually understand it??
of it that Rick didn't intend:


Much of what goes on today at the fore of AI can be viewed as
"expectation-simplified processing": frames, scripts, user modelling,
caching of all levels and flavors,...    Here are a few poins on the
continuum:
     After computing a value, cache it.  Note: "value" may be a number,
	a vector of numbers, an alphanumeric identifier, a vector of them...
     Synthesize a function capable of efficiently computing desired values,
        and cache that function.
     Synthesize a whole concept data structure (frame, script...) which
        contains relevant partial models of the recent situation.
There are more intermediate ones, but you see the progression from
static to dynamic, from scalar to descriptive.  In all cases, the
motivation for doing this storage is (i) speed up the recognition of
(the need to do) similar computations in the future, and (ii) speed up
or even eliminate those computations when you decide they ARE relevant
and you want the values they would return.  These are the two uses of
scripts, but also of cached constants (e.g., caching the value returned
by OWNED-BY, a function from CARS to PEOPLE). A better example: user models,
as they range from a single number, to a few numbers (e.g., in PARRY),
to a dynamic concept-modelling scheme.


OUTLINE:  

Introduction: 
 Model of sys. org. should include, perhaps, agenda-like control capabilities
** one reservation i have about this point and the subsequent developments
** related to it:  think cognitive!  lot's of computing, much parallelism,
** few agendas.  The notion of agenda should be abstracted from its current
** mechanism to its intrinsic purposes and properties,
** e.g., a tendency to respond first to those things that seem important,
**  demand (not just allow) attention, and for which ignoring them might
** lead to worse consequences than ignoring the competing alternatives
** Would all abstract computing machines for intelligent organisms
** use agendas to accomplish this?
** I think not, but all would still accomplish the basic functions.

 Model of intell. should stress the fact that knowledge acrretes, but only
   	occasionally shrinks (abandoning of a whole system of thought is
     	quite a rare event indeed.    --- ask Kuhn!)
		** good
 Model of computing can then, if we add the above 2 points, include this one:
  The half-frame problem
    In which the reasons supporting the tasks have the form "stiuation X is
  	not yet attained, but executing me will help you get there"
    	(e.g., facet f of concept C has only 2 entries)
    Then at task-choosing time, if a task's reasons are out-of-date, it is
 	most likely that that is because some of them are no longer valid,
	hence that the rating of the task will DECREASE
		** good
    So, all we need do is re-evaluate the top task, get a new LOWER number
	for its priority, reinsert it into the agenda, etc., and continue
	until the top task remains at the top.  The half-frame problem
	gets its name from  he fact that we do have to reevaluate some
   	tasks to see how the world has changed out from under them, BUT
	in general we need only do this for a small number of tasks to find
	the best one, not for ALL the tasks.
 Idea of Expec.: another point is that expectations allow us to set up
	frames in which to interpret the data, thereby making it v. quick.
		** okay
Abstraction
 As you know, I would like to rpesent heur. hier. as a kind of conceptual hier.
 I liked the idea of using bomber simulation, and will comment again on this.
	** i like the power of simulations to make our points too.
Caching
 Brian Smith (MIT grad student) gave me the general idea that the lisp EVAL
 function should, in general, have a second argument: the interpreter to use.
 One very special case of this is where EVAL takes as its second argument
 just a resource limit (time/space to expend); Bobrow and Norman discussed
 an idea like this a while back.  The continuum here stretches from the extreme
 of GETP (i,e, if a value is stored or cached there, fine, if not give up),
 to a more sophisticated inheritance search (accepting caches), 
 to the same thing but rejecting caches, to an even more complex deduction,
 perhaps with searching disk allowed, and finally to INduction (try to
 discover concepts relevant to answering the request).
** sure!
 The various headings Rick suggests are interrelated, and this may cause
 us some trouble later, esp. if we don't watch out for it.  E.g., the
 updating principles constrain and are constrained by the storage ones;
 When you store is constrained by Why; etc.
** sure, what'd'ya expect for a half-hour's work?

Expectation
 As stated above, generalize this enough to include stereotyping,
 of events (scripts), of situations (frames), of people (user models),...
**      nice idea

Cognitive econ. revisited
 This was quite good.
 What about the future, when processor costs drop way beolow storage, though?
** excellent question:  my answer--very little.  the economies we are
**talking about will still be true.  we are basically talking about ways
** to avoid computing things that are less desirable than alternatives.
** if you're intelligent, i presume there are always things you wish you
** could do if only you had time.  the faster you go, the behinder you get.

Phil's comments --- comments on them:

generally I ageed.  

The heuristic about "better talk about Hearts, not Bombers", is so
ivory tower as to be fit material for Sat. Night Live.  Unfortunately,
you are right that a good many AI researchers may feel that way.  We don't,
so the hell with them.  I did like your counterproposal about int'l.
terrorism, and I agree that we won't find ANY pro-terrorists at IJCAI.

Expectation-simplified processing IS awkward, but the concise replacements
generally have very negative connotations (e.g., stereotyping).
Expectation-focusing is not much better than the origianl term, in my opinion.


Let's get together and talk more about this.  One alternative to that
would be for someone, e.g. me, to flesh out the outline, essentially
start writing some prose.
** alternatively you could try to develop a detailed sentence outline,
** if you're able to work that way.  else, i agree, a draft is in order.

I am quite encouraged and think this will turn into a good paper.
** Me TOO.  Since thinking through these concepts, I've encountered numerous
** important opportunities for instantiating them and comprehending
** further their significance.
Doug




------- End of Forwarded Message


------- End of Forwarded Message




Doug-- the two papers are in the mail.  I presume you and Phil will
reconcile your seriality of authorship and the rest of us will
go alphabetically (that's why I changed from roth to hayes-roth).
I tried to marry isabella abacadella, but she cashed in on her
opportunitiy to run off with a certain smith.

	You may feel free to send future messages to me to me.

				Best wishes,

				   Rick


10 -- ************************
Mail from RAND-UNIX rcvd at 20-Nov-78 1508-PST
From: Klahr at Rand-Unix
Date: 20 Nov 1978 at 1508-PST
Message-Id: <[Rand-Unix]20-Nov-78 15:08:58.klahr>
To: lenat@aim, rick
cc: klahr
Subject: Cognitive Economy Paper


Doug and Rick,

     Title and authorship are fine with me.  With my luck, I'd probably
marry an yvonne young anyway.  But I don't need to get married just to
change my name -- how does Phil Aklahr sound?  After all, there is an
Akbar (Indian restaurant in the Marina), so Aklahr can't be far behind!

     Thoughts on catagories (keys) of cognitive economies:

     1.  Learning is a type of CE.  E.g., generalization can be viewed
	 as incorporating several instances into a single generalization.
	 Similarly for storing deductive chains as a single link (I
	 plan to expand this into a paper relative to my work in
	 deductive question-answering.)  Note this is only "one view"
	 (to quote a famous AI spokesman) of learning and certainly
	 "not the only one."

     2.  Inheritance should be an independent CE.  I assumed we would
	 talk about inheritance within Abstraction, but I think they
	 can and should be disjoint.  Abstraction can be used to
	 interpret at different levels of detail and as a planning
	 tool for focus of attention.  Inheritance saves storage and
	 updating at the expense of retrieval processing.

     3.  I agree with Rick's comment on the term "caching."  Our ideas
	 here are more general than that.  Perhaps we should distinguish
	 caching (storing values) and compiling (storing symbolic
	 expressions and compiled definitions).

Further thoughts as they come.

--Phil


12 -- ************************
Mail from RAND-UNIX rcvd at 20-Nov-78 1745-PST
From: Rick at Rand-Unix
Date: 20 Nov 1978 at 1745-PST
Message-Id: <[Rand-Unix]20-Nov-78 17:45:29.rick>
To: Lenat at SUMEX-AIM
cc: Rick at RAND-UNIX
Subject: Re: (Response to message)
In-reply-to: Your message of 20 Nov 1978 1559-PST.

Doug--
	It was because of the salutation that I was sure I was
an intended recipient.  But you'll note in my previous
msg to you, I incorporated the msg Phil received and forward to
me.  I think you just left me off the address list.  I've
assumed all along you wouldn't think I was actually upset.
But now I think I might leap off the pier.

		Bye,

			Rick


150 -- ************************
Mail from RAND-UNIX rcvd at 26-Dec-78 1108-PST
From: Rick at Rand-Unix
Date: 26 Dec 1978 at 1109-PST
To: lenat at Aim
cc: Rick at Rand-Unix, Klahr at Rand-Unix
Subject: joint paper

doug--
	january promises to be an incredibly hectic month.
both phil and i will be out of town different weeks, lots
of things will be happening, etc.  i encourage you to complete
the first draft asap.  how does your schedule look currently?

		best wishes,

			rick