Artificial Intelligence (AI) is a field of research that is both fascinating and mysterious, influenced by science fiction to such extent that many of us find it impossible to differentiate what is real and what is the fruit of the imagination of Hollywood writers and directors.
Artificial intelligence, watson and the final checkmate
1. Artificial
Intelligence,
Watson
and
the
final
checkmate
The
joining
of
genetics
to
bits,
the
essential
merger
March
03,
2011,
11:14
AM
By
Ricardo
Murer
B.S.
in
Computer
Science
(USP)
and
Master's
Degree
in
Communications
(USP).
Specialist
in
digital
strategy
and
new
technologies.
Follow@rdmurer
Artificial
Intelligence
(AI)
is
a
field
of
research
that
is
both
fascinating
and
mysterious,
influenced
by
science
fiction
to
such
extent
that
many
of
us
find
it
impossible
to
differentiate
what
is
real
and
what
is
the
fruit
of
the
imagination
of
Hollywood
writers
and
directors.
The
subject
recently
returned
to
the
media
when
IBM's
computer
"Watson”,
a
"Deep
Question
&Answer"
(DeepQA)
machine,
participated
in
the
American
TV
show
Jeopardy!,
beating
two
of
its
human
competitors.
The
show
is
exquisite
by
submitting
the
answers
to
participants
who
must
then
find
the
corresponding
question.
Watson
is
in
fact
a
combination
of
90
servers,
16
Terabytes
of
memory,
and
a
processing
capacity
of
180,000
Gigabytes
per
second!
The
most
beautiful
and
purest
parallel
brute
force
of
computing
today.
This
is
not
the
first
time
that
a
"deep"
beats
a
human
being
in
a
competition.
In
1997,
"Deep
Blue"
(you
must
remember),
an
IBM
chess
computer,
beat
the
then
champion
Garry
Kasparov.
Are
we
experiencing
the
onset
of
our
decline,
our
defeat
to
machines
smarter
than
we
are?
Fortunately,
not
yet.
In
both
cases,
we
are
referring
to
specialized
machines,
computers
able
to
perform
efficiently,
probabilistically
and
extremely
rapidly,
in
the
analysis
and
weighing
of
millions
of
possibilities,
and
through
a
combination
of
algorithms,
make
the
least
unfavorable
decision.
While
then
scientists
had
initially
intended
to
reproduce
the
laws
of
thought
and
create
a
machine
that
emulated
the
human
being,
the
approach
today
is
different.
That
is
because
the
objective
of
creating
a
machine
that
has
the
capacity
to
learn,
contemplate,
think
and
rationalize
has
proven
to
be
well
beyond
our
capacity.
The
complexity
of
such
a
task
stems
from
the
fact
that
intelligence
cannot
be
defined
and
understood
solely
by
one
field
of
knowledge
but
by
the
sum
of
all
fields
such
as
biology,
philosophy,
psychology
and
mathematics.
Among
the
new
approaches,
connectionism
attempts
to
draw
computer
processes
close
to
the
human
brain
through
"neural
networks",
while
the
symbolism
is
behind
specialist
systems.
AI
today
has
concentrated
many
of
its
efforts
into
simulating
human
cognitive
capacities
though
simulation
and
understanding
of
natural
language,
recognition
of
patterns
and
learning.
1
2. However,
we
are
still
far
from
creating
the
perfect
simulation
of
the
human
brain
and
eventually
something
that
can
be
deemed
the
result
of
intelligence,
even
if
by
something
of
a
rudimentary
intelligence.
Perhaps
it
is
because
the
computer,
the
one
that
counts,
or
the
ordinateur,
the
one
that
sorts
to
order
and
classifies,
remain,
in
spite
all
techno-‐scientific
advancements,
a
spectacular
calculating
and
counting
machine,
and
nothing
more.
Perhaps
because
genetics
is
yet
to
join
with
bits,
the
essential
merger
able
to
bring
light
to
the
first
thinking
android,
which
will
from
generation
to
generation
create
new
machines
able
to
evolve
from
simple
toolmakers
to
builders
of
complex
societies.
With
the
passing
of
time,
they
will
be
capable
of
simulating
not
just
intelligence,
but
also
other
human
capacities
such
as
aesthetic
feeling,
moral
judgment,
friendship
and
emotions,
converting
what
was
once
artificial
into
natural.
The
final
checkmate.
2