background image
259
IX TECHNOLOGY
Le
v
el of "ambient
intelligence"
Database
(str
uctur
ed data)
Information Base
(object-based
models)
Knowledge Base
(r
e-usable
knowledge)
Knowledge
Assistant
(inter
activ
e
knowledge)
Requir
ed quality
of service
"what y
ou put into it is
what y
ou get out of it"
Modelling of r
eal-w
orld
objects
Incr
easing ag
r
eement
for association (r
elation-
ship) modelling
Standar
dised association
modelling
Discr
ete kno
wledge
modules
Inter
operating single-
domain kno
wledge
modules
Inter
operating disparate
kno
wledge modules
Natural interaction via
str
uctur
ed interaction
language
Natural interaction via
fr
ee-text input
Natural interaction
thr
ough m
ulti-modal
interf
aces,
including
fr
ee speech
Requir
ed cor
e functionality
E-R Schema Modelling,
SQL Quer
ies
Unified Modelling Language
,
Object-
or
iented Design,Object quer
y language
Suppor
t for domain-specific patter
ns
and r
elationships
Standar
ds for gener
ic and domain-
specific patter
ns a and r
elationships
Kno
wledge Repr
esentation Languages
Domain-specific ontolo
g
i
es
Gener
ic
,
m
ulti-lingual ter
minolo
gy
r
esolving mechanisms
Interaction betw
een gener
ic kno
wledge
modules and domain-specific modules
(cf
.
Lenat,
"micr
o-theor
ies")
Belief models,
Agent interaction models
Context-sensiti
v
e
natural language
under
standing and generation
Context-sensiti
v
e
fr
ee-speech
generation
Applicable cor
e
technologies
Relational database
management systems
Object-or
iented databases
Fir
st-or
der lo
g
i
c
,
ter
mino-
lo
g
ical lo
g
ics intr
ospection
(quer
y the model and data)
Descr
iption lo
g
ics,
F
eatur
e lo
g
ics
F
eatur
e Lo
g
ics
Mix
ed-paradigm systems
(symbolic/subsymbolic)
Modal lo
g
ics,
temporal
lo
g
ics,
Multi-agent systems
Pr
ocess descr
iption lo
g
ics
Fuzzy
,
adapti
v
e
pr
ocesses
Scientific state of
the art (End 2001)
highly matur
e
f
airly matur
e
f
airly immatur
e
f
airly immatur
e
f
airly immatur
e
immatur
e
immatur
e
immatur
e
ve
ry
immatur
e
ve
ry
immatur
e
Scientific expectations for the
next 5-8 years
Marg
inal adv
ances in perfor
mance
Adv
ances in perfor
mance and suppor
t for
design and quer
y languages
Adv
ances in for
mal models that underlie
object-or
iented systems
Better under
standing of kno
wledge
modelling r
equir
ements
Better under
standing of domain model-
ling at the pr
ocess le
v
el;
o
v
er
coming of
kno
wn perfor
mance pr
ob
lems for
descr
iption/featur
e lo
g
ics
Br
eakthr
oughs in fuzzy matching
techniques for mix
ed-paradigm systems
W
orking pr
ototypes for significant
domain-specific pr
ob
lems
W
orking pr
ototypes for r
estr
icted
domains
Speculati
v
e
,
"to
y systems"