<s>
AutoTutor	B-Application
is	O
an	O
intelligent	O
tutoring	O
system	O
developed	O
by	O
researchers	O
at	O
the	O
Institute	O
for	O
Intelligent	O
Systems	O
at	O
the	O
University	O
of	O
Memphis	O
,	O
including	O
Arthur	O
C	O
.	O
Graesser	O
that	O
helps	O
students	O
learn	O
Newtonian	O
physics	O
,	O
computer	O
literacy	O
,	O
and	O
critical	O
thinking	O
topics	O
through	O
tutorial	O
dialogue	O
in	O
natural	O
language	O
.	O
</s>
<s>
AutoTutor	B-Application
differs	O
from	O
other	O
popular	O
intelligent	O
tutoring	O
systems	O
such	O
as	O
the	O
Cognitive	B-Application
Tutor	I-Application
,	O
in	O
that	O
it	O
focuses	O
on	O
natural	O
language	O
dialog	O
.	O
</s>
<s>
To	O
handle	O
this	O
input	O
,	O
AutoTutor	B-Application
uses	O
computational	O
linguistics	O
algorithms	O
including	O
latent	O
semantic	O
analysis	O
,	O
regular	B-Language
expression	I-Language
matching	O
,	O
and	O
speech	O
act	O
classifiers	O
.	O
</s>
<s>
In	O
addition	O
to	O
natural	O
language	O
input	O
,	O
AutoTutor	B-Application
can	O
also	O
accept	O
ad	O
hoc	O
events	O
such	O
as	O
mouse	O
clicks	O
,	O
learner	O
emotions	O
inferred	O
from	O
emotion	O
sensors	O
,	O
and	O
estimates	O
of	O
prior	O
knowledge	O
from	O
a	O
student	O
model	O
.	O
</s>
<s>
This	O
process	O
is	O
driven	O
by	O
a	O
"	O
script	O
"	O
that	O
includes	O
a	O
set	O
of	O
dialog-specific	O
production	B-Application
rules	I-Application
.	O
</s>
<s>
AutoTutor	B-Application
simulates	O
the	O
discourse	O
patterns	O
of	O
human	O
tutors	O
,	O
based	O
on	O
analysis	O
of	O
human-to-human	O
tutoring	O
sessions	O
and	O
theoretically-grounded	O
tutoring	O
strategies	O
based	O
on	O
cognitive	O
learning	O
principles	O
.	O
</s>
<s>
AutoTutor	B-Application
speaks	O
the	O
content	O
of	O
its	O
turns	O
through	O
an	O
animated	O
conversational	B-General_Concept
agent	I-General_Concept
with	O
a	O
speech	O
engine	O
,	O
some	O
facial	O
expressions	O
,	O
and	O
rudimentary	O
gestures	O
.	O
</s>
<s>
AutoTutor	B-Application
tracks	O
the	O
cognitive	O
states	O
of	O
the	O
learner	O
by	O
analyzing	O
the	O
content	O
of	O
the	O
dialogue	O
history	O
.	O
</s>
<s>
AutoTutor	B-Application
dynamically	O
selects	O
the	O
words	O
and	O
statements	O
in	O
each	O
conversational	O
turn	O
in	O
a	O
fashion	O
that	O
is	O
sensitive	O
to	O
what	O
the	O
learner	O
knows	O
.	O
</s>
<s>
Recent	O
versions	O
of	O
the	O
AutoTutor	B-Application
system	O
also	O
adapt	O
to	O
the	O
learner	O
's	O
emotional	O
states	O
in	O
addition	O
to	O
their	O
cognitive	O
states	O
.	O
</s>
<s>
AutoTutor	B-Application
has	O
shown	O
learning	O
gains	O
,	O
particularly	O
on	O
deep	O
reasoning	O
questions	O
,	O
in	O
over	O
a	O
dozen	O
experiments	O
on	O
college	O
students	O
for	O
topics	O
in	O
introductory	O
computer	O
literacy	O
and	O
conceptual	O
physics	O
.	O
</s>
<s>
Tests	O
of	O
AutoTutor	B-Application
have	O
produced	O
effect	B-Application
sizes	I-Application
with	O
a	O
mean	O
of	O
0.8	O
(	O
range	O
of	O
0.4	O
to	O
1.5	O
)	O
,	O
depending	O
on	O
the	O
learning	O
measure	O
,	O
the	O
comparison	O
condition	O
,	O
the	O
subject	O
matter	O
,	O
and	O
version	O
of	O
AutoTutor	B-Application
.	O
</s>
<s>
For	O
comparison	O
,	O
an	O
effect	B-Application
size	I-Application
of	O
1.0	O
would	O
be	O
roughly	O
equivalent	O
to	O
a	O
full	O
letter	O
grade	O
.	O
</s>
