/LMAR=0/XLINE=3/FONT#0=BAXL30/FONT#1=BAXM30/FONT#2=BASB30/FONT#3=SUB/FONT#4=SUP/FONT#5=BASL35/FONT#6=NGR25/FONT#7=MATH30/FONT#8=FIX25/FONT#9=GRKB30
␈↓ α∧␈↓␈↓ u1


␈↓ α∧␈↓␈↓↓The␈αFifth␈αGeneration␈α-␈αArtificial␈αIntelligence␈αand␈αJapan's␈αComputer␈αChallenge␈αto␈αthe␈αWorld␈↓␈α-␈αby
␈↓ α∧␈↓Edward Feigenbaum and Pamela McCorduck, Addison-Wesley Publishing Co.

␈↓ α∧␈↓␈↓ αTJapan␈α
has␈αreplaced␈α
the␈α
Soviet␈αUnion␈α
as␈α
the␈αworld's␈α
second␈α
place␈αindustrial␈α
power.␈α (Look␈α
at
␈↓ α∧␈↓the␈α∞globe␈α∞and␈α∞be␈α∞impressed).␈α∞ However,␈α
many␈α∞people,␈α∞Japanese␈α∞included,␈α∞consider␈α∞that␈α∞this␈α
has
␈↓ α∧␈↓relied␈αtoo␈αmuch␈αon␈αimported␈αscience␈αand␈αtechnology␈α-␈αtoo␈αmuch␈αfor␈αthe␈αrespect␈αof␈αthe␈αrest␈αof␈αthe
␈↓ α∧␈↓world,␈α⊃too␈α⊃much␈α⊃for␈α⊃Japanese␈α⊃self␈α∩respect,␈α⊃and␈α⊃too␈α⊃much␈α⊃for␈α⊃the␈α∩technological␈α⊃independence
␈↓ α∧␈↓needed␈α⊃for␈α⊃Japan␈α⊂to␈α⊃continue␈α⊃to␈α⊂advance␈α⊃at␈α⊃previous␈α⊂rates.␈α⊃ The␈α⊃Fifth␈α⊃Generation␈α⊂computer
␈↓ α∧␈↓project␈α
is␈α
one␈α∞Japanese␈α
attempt␈α
to␈α
break␈α∞out␈α
of␈α
the␈α
habit␈α∞of␈α
copying␈α
and␈α
generate␈α∞Japan's␈α
own
␈↓ α∧␈↓share of scientific and technological innovations.

␈↓ α∧␈↓␈↓ αTThe␈αidea␈αis␈αthat␈αthe␈α1990s␈αshould␈αsee␈αa␈αnew␈αgeneration␈αof␈αcomputers␈αbased␈αon␈α"knowledge
␈↓ α∧␈↓information␈α
processing"␈α
rather␈α
than␈α
"data␈αprocessing".␈α
 (The␈α
first␈α
generation␈α
computers␈αwere␈α
made
␈↓ α∧␈↓of␈α
vacuum␈αtubes,␈α
the␈α
second␈αof␈α
transistors␈α
and␈αthe␈α
third␈α
of␈αintegrated␈α
circuits.␈α"Fourth␈α
generation"
␈↓ α∧␈↓is␈α
an␈α
artefact␈α
of␈α
the␈α
trade␈α
press.␈α
 It␈α
was␈α
first␈α
used␈α
to␈α
refer␈α
to␈α
the␈α
IBM␈α
370␈α
though␈α
not␈α∞by␈α
IBM
␈↓ α∧␈↓which␈α∞wasn't␈α∂that␈α∞much␈α∞different␈α∂from␈α∞the␈α∞IBM␈α∂360,␈α∞but␈α∞has␈α∂also␈α∞been␈α∞used␈α∂to␈α∞refer␈α∂to␈α∞other
␈↓ α∧␈↓innovations).␈α∪ "Knowledge␈α∪information␈α∀processing"␈α∪is␈α∪a␈α∀vague␈α∪term␈α∪that␈α∀promises␈α∪important
␈↓ α∧␈↓advances in the direction of artificial intelligence but is non-committal about specific performance.

␈↓ α∧␈↓␈↓ αTThe␈α⊃Fifth␈α⊃Generation␈α∩Project␈α⊃(ICOT)␈α⊃is␈α∩the␈α⊃idea␈α⊃of␈α∩Kazuhiro␈α⊃Fuchi␈α⊃of␈α∩the␈α⊃Japanese
␈↓ α∧␈↓government's␈α≤Electro-Technical␈α≤Laboratory␈α≤-␈α≤ICOT,␈α≤while␈α≤supported␈α≤by␈α≤industry␈α≤and
␈↓ α∧␈↓government,␈α∀is␈α∀an␈α∀independent␈α∀institution.␈α∀ Fuchi␈α∀has␈α∀borrowed␈α∀about␈α∀forty␈α∃engineers␈α∀and
␈↓ α∧␈↓computer␈α
scientists,␈α
all␈α
under␈α
35,␈α
for␈α
periods␈α
of␈α
three␈α
years␈α
from␈α
the␈α
leading␈α
Japanese␈α
computer
␈↓ α∧␈↓companies.␈α Thus␈αthe␈αorganization␈α
and␈αmanagement␈αof␈αthe␈αproject␈α
is␈αas␈αinnovative␈αas␈α
one␈αcould
␈↓ α∧␈↓ask.␈α
 With␈αonly␈α
forty␈α
people,␈αthe␈α
project␈αis␈α
so␈α
far␈αa␈α
tiny␈αpart␈α
of␈α
the␈αtotal␈α
Japanese␈αcomputer␈α
effort.
␈↓ α∧␈↓It is scheduled to grow in subsequent phases.

␈↓ α∧␈↓␈↓ αTFrom␈α
its␈α
beginning␈α∞ICOT␈α
has␈α
been␈α∞hospitable␈α
to␈α
foreign␈α∞computer␈α
scientists.␈α
 It␈α∞has␈α
held
␈↓ α∧␈↓open conferences in English, publishes a journal in English and welcomes scientific visitors.

␈↓ α∧␈↓␈↓ αTIt␈α∀is␈α∃planned␈α∀to␈α∀take␈α∃about␈α∀ten␈α∀years.␈α∃ They␈α∀will␈α∀design␈α∃computers␈α∀based␈α∃on␈α∀"logic
␈↓ α∧␈↓programming",␈α
an␈α
invention␈αof␈α
Alain␈α
Colmerauer␈αof␈α
the␈α
University␈αof␈α
Marseilles␈α
in␈α
France␈αand
␈↓ α∧␈↓Robert␈α∞Kowalski␈α
of␈α∞Imperial␈α
College␈α∞in␈α
London,␈α∞and␈α
implemented␈α∞in␈α
a␈α∞computer␈α
programming
␈↓ α∧␈↓language␈αcalled␈α
Prolog.␈α They␈α
want␈αto␈α
use␈αadditional␈α
ideas␈αof␈α
"dataflow"␈αdeveloped␈α
at␈αM.I.T.␈α
 and
␈↓ α∧␈↓to␈α∞make␈α∞highly␈α∞parallel␈α
machines.␈α∞ Some␈α∞Japanese␈α∞university␈α
scientists␈α∞consider␈α∞that␈α∞the␈α
project
␈↓ α∧␈↓still has too much tendency to look to the West for scientific ideas.

␈↓ α∧␈↓␈↓ αTMaking␈α⊂parallel␈α⊂machines␈α⊂based␈α⊂on␈α⊂logic␈α⊂programming␈α⊂is␈α⊂a␈α⊂straightforward␈α∂engineering
␈↓ α∧␈↓task,␈α∂and␈α⊂there␈α∂is␈α∂little␈α⊂doubt␈α∂that␈α∂this␈α⊂part␈α∂of␈α∂the␈α⊂project␈α∂will␈α∂succeed.␈α⊂ The␈α∂grander␈α⊂goal␈α∂of
␈↓ α∧␈↓shifting␈α
the␈α
center␈α
of␈α
gravity␈α
of␈α
computer␈α
use␈α
to␈α
the␈α
intelligent␈α
processing␈α
of␈α
knowledge␈α
is␈αmore
␈↓ α∧␈↓doubtful␈α∩as␈α∩a␈α∩ten␈α∩year␈α∩effort.␈α∩ The␈α∪level␈α∩of␈α∩intelligence␈α∩to␈α∩be␈α∩achieved␈α∩is␈α∪ill-defined.␈α∩ The
␈↓ α∧␈↓applications␈α∩are␈α∪also␈α∩ill-defined.␈α∪ Some␈α∩of␈α∪the␈α∩goals,␈α∪such␈α∩as␈α∪common␈α∩sense␈α∪knowledge␈α∩and
␈↓ α∧␈↓reasoning ability require fundamental scientific discoveries which cannot be scheduled in advance.

␈↓ α∧␈↓␈↓ αTMy␈α⊂own␈α⊃scientific␈α⊂field␈α⊃is␈α⊂making␈α⊃computer␈α⊂programs␈α⊃with␈α⊂common␈α⊃sense,␈α⊂and␈α⊃when␈α⊂I
␈↓ α∧␈↓visited␈α
ICOT,␈αI␈α
asked␈αwho␈α
was␈αworking␈α
on␈αthe␈α
problem.␈α It␈α
was␈αdisappointing␈α
to␈αlearn␈α
that␈αthe
␈↓ α∧␈↓answer␈αwas␈α"no-one".␈α This␈αis␈αa␈αsubject␈αto␈αwhich␈αthe␈αJapanese␈αhave␈αmade␈αfew␈αcontributions,␈αand
␈↓ α∧␈↓it␈α
probably␈αisn't␈α
suited␈α
to␈αpeople␈α
borrowed␈α
from␈αcomputer␈α
companies␈α
for␈αthree␈α
years.␈α Therefore,␈α
I
␈↓ α∧␈↓can't␈α
be␈α
optimistic␈α
that␈α
this␈α
important␈α
part␈α
of␈α
the␈α
project␈α
goals␈α
will␈α
be␈α
achieved␈α
in␈α
the␈α
time␈αset.
␈↓ α∧␈↓␈↓ u2


␈↓ α∧␈↓Of␈αcourse,␈αany␈αprogress␈αis␈αlikely␈αto␈αbe␈αvery␈αimportant.␈α Moreover,␈αthis␈αproject␈αis␈αonly␈αa␈αsmall␈αpart
␈↓ α∧␈↓of Japan's computer research and development.  Their other work is also very competitive.

␈↓ α∧␈↓␈↓ αTThe␈α∪Fifth␈α∪Generation␈α∪Project␈α∩was␈α∪announced␈α∪at␈α∪a␈α∩time␈α∪when␈α∪the␈α∪Western␈α∩industrial
␈↓ α∧␈↓countries␈α∞were␈α∞ready␈α∞for␈α∞another␈α∞bout␈α∞of␈α∂viewing␈α∞with␈α∞alarm;␈α∞the␈α∞journalists␈α∞have␈α∞tired␈α∂of␈α∞the
␈↓ α∧␈↓"energy␈α⊗crisis"␈α⊗-␈α⊗not␈α∃that␈α⊗it␈α⊗has␈α⊗been␈α⊗solved.␈α∃ Even␈α⊗apart␈α⊗from␈α⊗the␈α⊗recession,␈α∃industrial
␈↓ α∧␈↓productivity␈α
has␈α
stagnated;␈αit␈α
has␈α
actually␈α
declined␈αin␈α
industries␈α
heavily␈α
affected␈αby␈α
environmental
␈↓ α∧␈↓and␈α∞safety␈α∞innovations.␈α∞ Meanwhile␈α
Japan␈α∞has␈α∞taken␈α∞the␈α
lead␈α∞in␈α∞automobile␈α∞production␈α∞and␈α
in
␈↓ α∧␈↓some other industries.

␈↓ α∧␈↓␈↓ αTAt␈αthe␈αsame␈αtime,␈α
artificial␈αintelligence␈αresearch␈αwas␈αgetting␈α
a␈αnew␈αround␈αof␈αpublicity␈α
which
␈↓ α∧␈↓seems␈α∞to␈α∞go␈α∂in␈α∞a␈α∞seven␈α∂year␈α∞cycle.␈α∞ For␈α∂a␈α∞while␈α∞every␈α∞editor␈α∂wants␈α∞a␈α∞story␈α∂on␈α∞AI␈α∞and␈α∂the␈α∞free
␈↓ α∧␈↓lancers␈α
oblige,␈α
and␈α
then␈α
suddenly␈α
the␈α
editors␈α
get␈α
tired␈α
of␈α
it.␈α
 This␈α
round␈α
of␈α
publicity␈α
has␈α
more␈α
new
␈↓ α∧␈↓facts␈α
behind␈αit␈α
than␈αbefore,␈α
because␈αexpert␈α
systems␈αare␈α
beginning␈αto␈α
achieve␈αpractical␈α
results,␈αi.e.
␈↓ α∧␈↓results that companies will pay money for.

␈↓ α∧␈↓␈↓ αTTherefore,␈α∩the␈α⊃Fifth␈α∩Generation␈α∩Project␈α⊃has␈α∩received␈α⊃enormous␈α∩publicity,␈α∩and␈α⊃Western
␈↓ α∧␈↓computer␈α⊗scientists␈α⊗have␈α⊗taken␈α↔it␈α⊗as␈α⊗an␈α⊗occasion␈α↔for␈α⊗spurring␈α⊗their␈α⊗colleagues␈α↔and␈α⊗their
␈↓ α∧␈↓governments.␈α Apocalyptic␈αlanguage␈αis␈αused␈αthat␈αsuggests␈αthat␈αthere␈αis␈αa␈αbattle␈αto␈αthe␈αdeath␈α-␈αonly
␈↓ α∧␈↓one␈α
computer␈α
industry␈α
can␈α
survive,␈αtheirs␈α
or␈α
ours.␈α
 Either␈α
we␈αsolve␈α
all␈α
the␈α
problems␈α
of␈αartificial
␈↓ α∧␈↓intelligence right away or they walk all over us.

␈↓ α∧␈↓␈↓ αTEdward␈α⊂Feigenbaum␈α⊂is␈α⊃the␈α⊂leader␈α⊂of␈α⊃one␈α⊂of␈α⊂the␈α⊂major␈α⊃groups␈α⊂that␈α⊂has␈α⊃achieved␈α⊂these
␈↓ α∧␈↓results␈α⊂with␈α⊂programs␈α⊂applicable␈α∂to␈α⊂chemistry␈α⊂and␈α⊂medicine.␈α⊂ He␈α∂is␈α⊂also␈α⊂one␈α⊂of␈α⊂the␈α∂American
␈↓ α∧␈↓computer␈α⊃scientists␈α∩with␈α⊃extensive␈α⊃Japanese␈α∩contacts␈α⊃and␈α⊃extensive␈α∩interaction␈α⊃with␈α∩the␈α⊃Fifth
␈↓ α∧␈↓Generation Project.

␈↓ α∧␈↓␈↓ αTPamela␈α∞McCorduck␈α
is␈α∞a␈α∞science␈α
writer␈α∞with␈α
a␈α∞previous␈α∞book␈α
␈↓↓Machines␈α∞Who␈α∞Think␈↓␈α
about
␈↓ α∧␈↓the history of artificial intelligence research.

␈↓ α∧␈↓␈↓ αTFeigenbaum and McCorduck make two main points.

␈↓ α∧␈↓␈↓ αTFirst, knowledge engineering will dominate computing by the 1990s.

␈↓ α∧␈↓␈↓ αTSecond,␈αAmerica␈αis␈αin␈αdeep␈αtrouble␈αif␈αwe␈αdon't␈αorganize␈αa␈αsystematic␈αeffort␈αto␈αcompete␈αwith
␈↓ α∧␈↓the Japanese in this area.

␈↓ α∧␈↓␈↓ αTI␈αagree␈αwith␈αthem␈αthat␈αknowledge␈αengineering␈αwill␈αincrease␈αin␈αimportance,␈αbut␈αI␈αthink␈αthat
␈↓ α∧␈↓many␈α∂of␈α∂its␈α∂goals␈α∂will␈α∂require␈α∂fundamental␈α∂scientific␈α∂advances;␈α∂these␈α∂cannot␈α∂be␈α∂scheduled␈α∂to␈α∂a
␈↓ α∧␈↓fixed␈α⊃time␈α∩frame.␈α⊃ What's␈α∩worse,␈α⊃even␈α∩in␈α⊃the␈α∩United␈α⊃States␈α∩and␈α⊃Britain,␈α∩the␈α⊃hope␈α∩of␈α⊃quick
␈↓ α∧␈↓applications␈α∪has␈α∪lured␈α∪too␈α∪many␈α∪students␈α∪away␈α∪from␈α∪basic␈α∪research.␈α∪ I␈α∪also␈α∪agree␈α∪that␈α∪our
␈↓ α∧␈↓industrial␈α
system␈αhas␈α
serious␈αweaknesses␈α
some␈αof␈α
which␈αthe␈α
Japanese␈αhave␈α
avoided.␈α If␈α
we␈αwere␈α
to
␈↓ α∧␈↓match␈α⊂their␈α⊂forty␈α⊃engineer␈α⊂project␈α⊂according␈α⊃to␈α⊂output␈α⊂of␈α⊃our␈α⊂educational␈α⊂system,␈α⊃our␈α⊂project
␈↓ α∧␈↓would have twenty engineers and twenty lawyers.

␈↓ α∧␈↓␈↓ αTThe␈αauthors␈αare␈α
properly␈αcautious␈αabout␈αwhat␈α
kind␈αof␈αan␈αAmerican␈α
project␈αis␈αcalled␈αfor.␈α
 It
␈↓ α∧␈↓simply␈αcannot␈αbe␈αan␈αApollo-style␈αproject,␈αbecause␈αthat␈αdepended␈αon␈αhaving␈αa␈αrather␈αprecise␈αplan
␈↓ α∧␈↓in␈α∂the␈α∂beginning␈α∂that␈α⊂could␈α∂see␈α∂all␈α∂the␈α⊂way␈α∂to␈α∂the␈α∂end␈α∂and␈α⊂did␈α∂not␈α∂depend␈α∂on␈α⊂new␈α∂scientific
␈↓ α∧␈↓discoveries.␈α Activities␈αthat␈αwere␈αpart␈αof␈αthe␈αplan␈αwere␈αpushed,␈αand␈αeverything␈αthat␈αwas␈α
not␈αpart
␈↓ α∧␈↓␈↓ u3


␈↓ α∧␈↓of␈α
it␈αwas␈α
ruthlessly␈αtrimmed.␈α
 This␈α
would␈αbe␈α
disastrous␈αwhen␈α
we␈α
cannot␈αpredict␈α
what␈αresearch␈α
will
␈↓ α∧␈↓be␈αrelevant␈αto␈αthe␈αgoal.␈α My␈αown␈αview␈αis␈αrather␈αconventional.␈α We␈αwill␈αmake␈αthe␈αmost␈αprogress␈αif
␈↓ α∧␈↓there␈αis␈α
money␈αto␈αsupport␈α
unsolicited␈αproposals,␈αi.e.␈α
where␈αthe␈αresearch␈α
defines␈αthe␈αgoals␈α
and␈αthe
␈↓ α∧␈↓funders␈α∪decide␈α∪whether␈α∩he␈α∪and␈α∪his␈α∪project␈α∩are␈α∪better␈α∪than␈α∩his␈α∪competition.␈α∪ We␈α∪also␈α∩need
␈↓ α∧␈↓postdoctoral fellowships to give the best new PhDs a chance to come up with conceptual advances.