To wydarzenie nei przynioslo tyle sensacji co turniej szachosy meidzy
komputerem i zlowiekiem, ale tez nei przeszko pzrez media
neizauwazone: komputer wygral z czlowiekiem w teleturnieju zwanum
Jeopardy.

Okazuje sie ze "inteligentna" czesc programu byla napisana w Prologu

The final round in the epic man vs. machine battle that's been playing
out on Jeopardy! all week was fought tonight. The winner: Watson,
IBM's supercomputer, who soundly defeated flesh-and-blood opponents
Ken Jennings and Brad Rutter, ending a three-night tournament and
beginning a long period of social and technological commentators
trying to figure out what it all meant.

http://www.pcmag.com/article2/0,2817,2380489,00.asp

This what is interesting, is that the "intelligent" part of the
program was done in Prolog:

Several newspapers published
internal details on the Watson computer, some of them unveiling that
Watson used Prolog for some of the natural language processing of the
Jeopardy! clues.
...
The Prolog language is
the most adequate language for pattern matching (unification) and
problems that involve a depth-first search and backtracking (in
imperative languages like C and Java such pattern DFS problems would
be implemented by a visitor pattern or a (really giant) cascade of IF-
THEN-ELSE statements or a CASE unification (if possible)) because it
is a lot tidier to write all the potential patterns as Prolog clauses
of the form "head if body conjunction" and query Prolog for new
annotations (information about the clue). Although simple, the Prolog
language is very expressive allowing constructs difficult to represent
in other languages (such as, recursive rules to represent reachability
in parse trees, the use of negation-as-failure to check the absence of
annotations). It is also easy to manage data structures of unknown
depth in Prolog due to functors and the operation of unification (very
useful for representing sequences, parse trees, built-ins for list
handling).
...
Trying to implement all these requirements/features in
another language tends to be either a mess OR a reproduction of Prolog
(the cleanest and most general solution probably involves implementing
top-down evaluation and unification).

http://www.rhinocerus.net/forum/lang-prolog/662373-i
bm-watsons-win-jeopardy-also-success-story-prolog.ht
ml

A.L.