/LMAR=0/XLINE=3/FONT#0=BASL30/FONT#1=BASI30/FONT#2=BASB30/FONT#5=SUP/FONT#3=STA200/FONT#4=NGB25
␈↓ ↓H␈↓␈↓βS␈↓∧ Artificial Intelligence Laboratory, STANFORD UNIVERSITY, Stanford, California 94305
␈↓ ↓H␈↓∧Telephone 415 497-4971␈↓ 

August 6, 1975




␈↓ ↓H␈↓Professor J. Gelfand
␈↓ ↓H␈↓Council of Cybernetics
␈↓ ↓H␈↓USSR Academy of Sciences
␈↓ ↓H␈↓111333, 40 Vavilov Street
␈↓ ↓H␈↓Moscow V-333, USSR

␈↓ ↓H␈↓Dear Professor Gelfand:

␈↓ ↓H␈↓        Thank␈α∪you␈α∩for␈α∪your␈α∩letter␈α∪about␈α∩the␈α∪panel␈α∩on␈α∪"Mathematics,␈α∩Cybernetics␈α∪and␈α∪AI"␈α∩at
␈↓ ↓H␈↓4IJCAI.␈α
 As␈αyou␈α
know,␈α
I␈αam␈α
replacing␈α
Prof.␈α Jack␈α
Minker␈αas␈α
co-chairman␈α
of␈αthe␈α
panel␈α
and␈αdid
␈↓ ↓H␈↓not␈αmake␈αthe␈αoriginal␈αarrangements,␈αbut␈αhere␈αare␈αmy␈αideas␈αof␈αhow␈αto␈αproceed␈αfrom␈αwhere␈αhe␈αleft
␈↓ ↓H␈↓o≥.

␈↓ ↓H␈↓        Both␈α⊂␈↓↓cybernetics␈↓␈α⊂as␈α⊂orignally␈α⊂conceived␈α∂by␈α⊂Norbert␈α⊂Wiener␈α⊂and␈α⊂␈↓↓arti≡cial␈α⊂intelligence␈↓␈α∂are
␈↓ ↓H␈↓concerned␈α∞with␈α
goal-seeking␈α∞systems.␈α
 However,␈α∞cybernetics,␈α
both␈α∞as␈α
initially␈α∞conceived␈α
and␈α∞as␈α
it
␈↓ ↓H␈↓subsequently␈αdeveloped,␈αis␈α
to␈αa␈αlarge␈α
extent␈αconcerned␈αwith␈αthe␈α
case␈αin␈αwhich␈α
the␈αstate␈αspace␈αis␈α
R␈↓¬n␈↓
␈↓ ↓H␈↓and␈α
provided␈αwith␈α
a␈α
metric␈αfor␈α
evaluating␈α
the␈αdistance␈α
of␈α
a␈αstate␈α
from␈α
the␈αgoal␈α
state␈αtogether␈α
with
␈↓ ↓H␈↓a␈α
rule␈α
that␈α
gives␈α
the␈α
e≥ects␈α
of␈α
actions.␈α
 It␈α
seems␈α
to␈α
me␈α
that␈α
the␈α
emphasis␈α
in␈α
cybernetics␈α
has␈α
been␈α
on
␈↓ ↓H␈↓those␈α∪cases␈α∩in␈α∪which␈α∩optimal␈α∪strategies␈α∪can␈α∩be␈α∪determined␈α∩mathematically␈α∪and␈α∪about␈α∩which
␈↓ ↓H␈↓theorems␈α∂can␈α∂be␈α∂proved.␈α∂ As␈α∂the␈α⊂mathematics␈α∂has␈α∂improved,␈α∂the␈α∂cases␈α∂that␈α∂have␈α⊂been␈α∂treated
␈↓ ↓H␈↓have␈αexpanded.␈α Besides␈αthis␈αnarrow␈αsense␈αof␈αcybernetics,␈αinstitutions␈αwith␈αthe␈αword␈α"cybernetics"
␈↓ ↓H␈↓in their name have undertaken a variety of activities some of which they also call "cybernetics".

␈↓ ↓H␈↓        Arti≡cial␈α∞intelligence␈α
is␈α∞concerned␈α
with␈α∞the␈α
activities␈α∞involved␈α
in␈α∞intelligent␈α∞behavior,␈α
and
␈↓ ↓H␈↓the␈α
main␈αinitial␈α
activity␈αwas␈α
the␈αsearch␈α
of␈αdiscrete␈α
spaces␈αsuch␈α
as␈αmove␈α
trees␈αand␈α
proof␈αtrees.␈α
 The
␈↓ ↓H␈↓emphasis␈α∂has␈α∂been␈α∞rather␈α∂empirical,␈α∂and␈α∞the␈α∂algorithms␈α∂studied␈α∞have␈α∂rarely␈α∂lent␈α∂themselves␈α∞to
␈↓ ↓H␈↓mathematical␈α∂treatment.␈α⊂ It␈α∂seems␈α⊂quite␈α∂clear␈α⊂to␈α∂workers␈α⊂in␈α∂AI␈α⊂that␈α∂the␈α⊂mathematical␈α∂methods
␈↓ ↓H␈↓developed␈α∂in␈α∂cybernetics␈α∂have␈α∂a␈α⊂limited␈α∂domain␈α∂of␈α∂application,␈α∂especially␈α∂where␈α⊂the␈α∂intelligent
␈↓ ↓H␈↓system must itself conceptualize the problem, but also in simpler cases.

␈↓ ↓H␈↓        Nevertheless,␈αthere␈αis␈αa␈αcertain␈αoverlap␈αbetween␈αactivities␈αcalled␈αcybernetics␈αand␈αthose␈αcalled
␈↓ ↓H␈↓arti≡cial␈α_intelligence.␈α_ In␈α_particular,␈α_some␈α↔kinds␈α_of␈α_search␈α_procedures␈α_have␈α_been␈α↔studied
␈↓ ↓H␈↓mathematically,␈α∪and␈α∀some␈α∪methods␈α∪such␈α∀as␈α∪dynamic␈α∪programming␈α∀have␈α∪been␈α∀suggested␈α∪for
␈↓ ↓H␈↓application to AI type problems.

␈↓ ↓H␈↓        In␈α∩my␈α∪view,␈α∩it␈α∩would␈α∪be␈α∩interesting␈α∩for␈α∪the␈α∩panel␈α∩members␈α∪to␈α∩discuss␈α∩what␈α∪kinds␈α∩of
␈↓ ↓H␈↓problems␈α∞lend␈α∞themselves␈α∞to␈α∞the␈α∞kind␈α∞of␈α∞mathematics␈α∞developed␈α∞in␈α∞connection␈α∂with␈α∞cybernetics.
␈↓ ↓H␈↓Perhaps␈α∂there␈α∂might␈α∂be␈α∂a␈α∞survey␈α∂of␈α∂these␈α∂methods␈α∂from␈α∞the␈α∂cybernetics␈α∂side,␈α∂and␈α∂a␈α∂survey␈α∞of
␈↓ ↓H␈↓problems␈α
from␈α
the␈α
AI␈α
side.␈α
 This␈α
might␈α
then␈αresult␈α
in␈α
a␈α
fairly␈α
lively␈α
discussion␈α
of␈α
the␈αprospects
␈↓ ↓H␈↓for the further extension of mathematical methods.

␈↓ ↓H␈↓        Perhaps␈αthere␈αwill␈αbe␈αsubstantial␈αdisagreements␈αbetween␈αworkers␈αin␈αthe␈αtwo␈αdomains␈αabout
␈↓ ↓H␈↓αProfessor J. Gelfand␈↓ ¬hAugust 6, 1975␈↓ 
nPage 2␈↓ 

␈↓ ↓H␈↓what␈αproblems␈αlend␈αthemselves␈αto␈αmathematical␈αmethods.␈α For␈αexample,␈αI␈αremember␈αthat␈αRichard
␈↓ ↓H␈↓Bellman␈α
was␈α
of␈α
the␈α
opinion␈α
that␈α
dynamic␈α
programming␈α
methods␈α
would␈α
work␈α
for␈α
chess,␈α
and␈αwe␈α
AI
␈↓ ↓H␈↓workers␈α
disagreed,␈αbut␈α
I␈αdon't␈α
know␈αanyone␈α
who␈α
will␈αbe␈α
at␈αTblisi␈α
who␈αwould␈α
take␈αBellman's␈α
point
␈↓ ↓H␈↓of view.

␈↓ ↓H␈↓        As␈αto␈α
participants,␈αBertram␈αRaphael␈α
and␈αAlexander␈α
Lerner␈αand␈αI␈α
were␈αalready␈α
invited␈αby
␈↓ ↓H␈↓Minker␈α
and␈α∞con≡rmed␈α
by␈α
the␈α∞the␈α
conference␈α∞chairman␈α
Professor␈α
Winston.␈α∞ Besides␈α
this,␈α∞I␈α
have
␈↓ ↓H␈↓asked␈αMax␈αClowes␈αto␈αtake␈αpart.␈α As␈αI␈αrecall,␈αAndrei␈αErshov␈αand␈αVictor␈αBriabrin␈αwere␈αalso␈αgoing
␈↓ ↓H␈↓to␈α⊃take␈α⊃part,␈α⊂but␈α⊃I␈α⊃have␈α⊂not␈α⊃been␈α⊃in␈α⊂touch␈α⊃with␈α⊃them.␈α⊂ I␈α⊃suppose␈α⊃that␈α⊂you␈α⊃will␈α⊃verify␈α⊂their
␈↓ ↓H␈↓intentions and invite whomever else you think will make a good contribution.

␈↓ ↓H␈↓        As␈α∞you␈α∞may␈α
already␈α∞know␈α∞and␈α∞can␈α
see␈α∞from␈α∞the␈α∞enclosures,␈α
there␈α∞has␈α∞been␈α∞some␈α
concern
␈↓ ↓H␈↓about␈αthe␈αpossibility␈αthat␈αAlexander␈αLerner␈αmight␈αnot␈αbe␈αallowed␈αto␈αparticipate␈αin␈αthe␈αpanel.␈α In
␈↓ ↓H␈↓order␈α∀to␈α∃keep␈α∀matters␈α∀clear,␈α∃I␈α∀will␈α∀take␈α∃the␈α∀responsibility␈α∀for␈α∃continuing␈α∀his␈α∃invitation␈α∀to
␈↓ ↓H␈↓participate␈α
as␈αProfessor␈α
Winston␈αhas␈α
asked␈αme␈α
to␈αdo.␈α
 This␈αis␈α
entirely␈αregular,␈α
because␈α
IJCAI␈αis
␈↓ ↓H␈↓an␈α
organization␈α
of␈α
individuals␈α
and␈α
not␈αof␈α
national␈α
organizations,␈α
and␈α
the␈α
Program␈αChairman,␈α
Pat
␈↓ ↓H␈↓Winston,␈α∞was␈α∞given␈α∞complete␈α∞authority␈α∞over␈α∞the␈α∞program␈α∞at␈α∞the␈α∞last␈α∞Conference␈α∞at␈α∞Stanford␈α∞in
␈↓ ↓H␈↓1973.  (The previous chairman, Nils Nilsson had complete authority over that meeting).

␈↓ ↓H␈↓        Along␈α⊃with␈α⊂Lerner␈α⊃and␈α⊂Raphael,␈α⊃I␈α⊂am␈α⊃inviting␈α⊂Max␈α⊃Clowes␈α⊂to␈α⊃participate,␈α⊃because␈α⊂he
␈↓ ↓H␈↓knows␈α∞quite␈α∞a␈α∞bit␈α∞about␈α∂both␈α∞arti≡cial␈α∞intelligence␈α∞and␈α∞work␈α∂that␈α∞has␈α∞gone␈α∞under␈α∞the␈α∂name␈α∞of
␈↓ ↓H␈↓cybernetics.

␈↓ ↓H␈↓        I␈αam␈αlooking␈αforward␈αto␈αmeeting␈αwith␈αyou␈αin␈αTblisi␈αor␈αin␈αMoscow␈αwhere␈αI␈αam␈αexpecting␈αto
␈↓ ↓H␈↓arrive on Sunday August 31.


␈↓ ↓H␈↓Sincerely,



␈↓ ↓H␈↓John McCarthy
␈↓ ↓H␈↓Director
␈↓ ↓H␈↓Professor of Computer Science