gennari vittorini-terence-ai*ia-2014
TRANSCRIPT
Improving*Children’s*Reading*Comprehension
Arti6icial*Intelligence
for
Rosella*Gennari
http://www.inf.unibz.it/~gennari
Pierpaolo*Vittorini*http://vittorini.univaq.it
www.terenceproject.eu
ONCE%UPON%A%TIME…
[…]%After%two%days%of%hot%sun%and%of%swimming%in%the%lake,%today%it%was%the%big%match%for%Alessio,%the%last%game%of%the%season%in%his%home%town,%l’Aquila.%There%was%only%a%minute%of%the%game%left.%The%crowd%watched%in%silence%as%Alessio%took%the%penalty%shot…%%And%the%crowd%of%l’Aquila%cheered%and%roared!
ONCE%UPON%A%TIME…
What%does%Alessio%play?
[…]%After%two%days%of%hot%sun%and%of%swimming%in%the%lake,%today%it%was%the%big%match%for%Alessio,%the%last%game%of%the%season%in%his%home%town,%l’Aquila.%There%was%only%a%minute%of%the%game%left.%The%crowd%watched%in%silence%as%Alessio%took%the%penalty%shot…%%And%the%crowd%of%l’Aquila%cheered%and%roared!
When%does%the%game%take%place?%
Who%wins?
ONCE%UPON%A%TIME…
[…]%After%two%days%of%hot*sun*and*of*swimming*in*the*lake,%today%it%was%the%big%match%for%Alessio,%the%last%game%of%the%season%in%his%home%town,%l’Aquila.%There%was%only%a%minute%of%the%game%left.%The%crowd%watched%in%silence%as%Alessio%took%the%penalty%shot…%%And%the%crowd%of%l’Aquila%cheered%and%roared!
What%does%Alessio%play?
When%does%the%game%take%place?%
Who%scores?
ONCE%UPON%A%TIME…
[…]%After%two%days%of%hot%sun%and%of%swimming%in%the%lake,%today%it%was%the%big%match%for%Alessio,*the*last*game*of*the*season*in*his*home*town,*l’Aquila.%There%was%only%a%minute%of%the%game%left.%The*crowd*watched*in*silence*as*Alessio*took*the*penalty*shot…**And*the*crowd*of*l’Aquila*cheered*and*roared!
What%does%Alessio%play?
When%does%the%game%take%place?%
Who%scores?
ONCE%UPON%A%TIME…
[…]%After%two%days%of%hot%sun%and%of%swimming%in%the%lake,%today%it%was%the%big%match%for%Alessio,%the*last*game*of*the*season*in*his*home*town,*l’Aquila.*There*was*only*a*minute*of*the*game*left.*The*crowd*watched*in*silence*as*Alessio*took*the*penalty*shot…**And*the*crowd*of*l’Aquila*cheered*and*roared!
archivio.riparteilfuturo.it
What%does%Alessio%play?
When%does%the%game%take%place?%
Who%scores?
THE%PROBLEM
‣More%than%10%%of%primary%school%children,%older%than%8,%are%diagnosed%with%deep%text%comprehension%problems,%e.g.,%they%fail%answering%
when%does%the%game%take%place?%%
‣They%are%referred%to%as%poor*comprehenders
?
THE%PROBLEM
‣More%than%10%%of%primary%school%children,%older%than%8,%are%diagnosed%with%deep%text%comprehension%problems,%e.g.,%they%fail%answering%
when%does%the%game%take%place?%%
‣They%are%referred%to%as%poor*comprehenders
‣Can%technology%help%them?
?
OUTLINE
1. Technology%enhanced%learning%
2. The%TERENCE%case%study%
3. Game%over%
AI**TEL
TECHNOLOGY%ENHANCED%LEARNING%(TEL)="technology2for2the2growth"of"a"learning"experience
AI**TEL
TECHNOLOGY%ENHANCED%LEARNING%(TEL)="technology2for2the2growth"of"a"learning"experience
innaffiatoi
o
AI**TEL
TECHNOLOGY%ENHANCED%LEARNING%(TEL)="technology2for2the2growth"of"a"learning"experience
innaffiatoi
o
ARTIFICIAL INTELLIGENCE (AI)
AI**TEL
TECHNOLOGY%ENHANCED%LEARNING%(TEL)="technology2for2the2growth"of"a"learning"experience
innaffiatoi
o
How%can%we%%design%and%evaluate%%AI%products%that%support%%the%growth%of%their%users’%learning*experience?%
ARTIFICIAL INTELLIGENCE (AI)
TEL%4%LEARNING%EXPERIENCE
Montessori:2adequate"tasks2that2come2in2a2prepared"environment2designed2on2top2of2the2learner2characteristics"can2effectively2support2learning
innaffiatoi
o
ousability%of%%learning%products
TEL%4%LEARNING%EXPERIENCE
Montessori:2adequate"tasks2that2come2in2a2prepared"environment2designed2on2top2of2the2learner2characteristics"can2effectively2support2learning
innaffiatoi
o
ousability%of%%learning%products
opedagogical%effectiveness%%of%products%for%learning*growth
TEL%4%LEARNING%EXPERIENCE
Montessori:2adequate"tasks2that2come2in2a2prepared"environment2designed2on2top2of2the2learner2characteristics"can2effectively2support2learning
innaffiatoi
o
OUTLINE1. Student%centred%learning%and%technology%
enhanced%learning%
2. The%TERENCE%case%study%3. Game%over
THE%TERENCE%ADAPTIVE%SOLUTION
w"h"o""a"r"e""y"o"u"?
www.terenceproject.eu
THE%TERENCE%ADAPTIVE%SOLUTION
w"h"o""a"r"e""y"o"u"?
a"d"e"q"u"a"t"e"b"o"o"k""o"f"""s"t"o"r"i"e"s
www.terenceproject.eu
THE%TERENCE%ADAPTIVE%SOLUTION
w"h"o""a"r"e""y"o"u"?
a"d"e"q"u"a"t"e"b"o"o"k""o"f"""s"t"o"r"i"e"s
a"d"e"q"u"a"t"e"s"m"a"r"t"g"a"m"e"s
www.terenceproject.eu
THE%TERENCE%ADAPTIVE%SOLUTION
w"h"o""a"r"e""y"o"u"?
a"d"e"q"u"a"t"e"b"o"o"k""o"f"""s"t"o"r"i"e"s
r"e"w"a"r"da"d"e"q"u"a"t"e"s"m"a"r"t"g"a"m"e"s
www.terenceproject.eu
THE%TERENCE%ADAPTIVE%SOLUTION
a"d"e"q"u"a"t"e"b"o"o"k""o"f"""s"t"o"r"i"e"s
s"i"g"n""i"n
a"d"e"q"u"a"t"e"s"m"a"r"t"g"a"m"e"s
r"e"w"a"r"d
www.terenceproject.eu
Requirements Prototypes Analytic + small ev. Int. prod. Large ev. Fin. prod.
HOW%DID%WE%GET%TO%THE%TERENCE%SOLUTION
www.terenceproject.eu
Requirements Prototypes Analytic + small ev. Int. prod. Large ev. Fin. prod.
HOW%DID%WE%GET%TO%THE%TERENCE%SOLUTION
www.terenceproject.eu
Requirements Prototypes Analytic + small ev. Int. prod. Large ev. Fin. prod.
HOW%DID%WE%GET%TO%THE%TERENCE%SOLUTION
www.terenceproject.eu
Requirements Prototypes Analytic + small ev. Int. prod. Large ev. Fin. prod.
HOW%DID%WE%GET%TO%THE%TERENCE%SOLUTION
www.terenceproject.eu
Requirements Prototypes Analytic + small ev. Int. prod. Large ev. Fin. prod.
HOW%DID%WE%GET%TO%THE%TERENCE%SOLUTION
www.terenceproject.eu
problem: 256 stories,
each with ~12 games
Smart*game*design
www.terenceproject.eu
how can we automatise the
development of smart games via AI (and, hopefully, be efficient)?
Smart*game*design
www.terenceproject.eu
story
ALS LayerGUI Layer
Learner
EducatorExpert
Learner GUI
Expert GUI
Persistence Layer
OpenRDF
UserManager
OpenRDF
StoryManager
OpenRDF
GameManager
OpenRDF
VisualisationManagerillustrations
NPL
Reasoner
AdaptiveEngine
Visualisation
������������
Reasoning Module
AnnotationModule
VisualisationModule
SemiRautomated*generation
www.terenceproject.eu
story
annotations
ALS LayerGUI Layer
Learner
EducatorExpert
Learner GUI
Expert GUI
Persistence Layer
OpenRDF
UserManager
OpenRDF
StoryManager
OpenRDF
GameManager
OpenRDF
VisualisationManagerillustrations
NPL
Reasoner
AdaptiveEngine
Visualisation
������������
Reasoning Module
AnnotationModule
VisualisationModule
SemiRautomated*generation
www.terenceproject.eu
enriched annotations
story
annotations
ALS LayerGUI Layer
Learner
EducatorExpert
Learner GUI
Expert GUI
Persistence Layer
OpenRDF
UserManager
OpenRDF
StoryManager
OpenRDF
GameManager
OpenRDF
VisualisationManagerillustrations
NPL
Reasoner
AdaptiveEngine
Visualisation
������������
Reasoning Module
AnnotationModule
VisualisationModule
SemiRautomated*generation
www.terenceproject.eu
enriched annotations
story
annotations
SemiRautomated*generation
ALS LayerGUI Layer
Learner
EducatorExpert
Learner GUI
Expert GUI
Persistence Layer
OpenRDF
UserManager
OpenRDF
StoryManager
OpenRDF
GameManager
OpenRDF
VisualisationManagerillustrations
NPL
Reasoner
AdaptiveEngine
Visualisation
������������
Reasoning Module
AnnotationModule
VisualisationModule
www.terenceproject.eu
enriched annotations
story
annotations
SemiRautomated*generation
ALS LayerGUI Layer
Learner
EducatorExpert
Learner GUI
Expert GUI
Persistence Layer
OpenRDF
UserManager
OpenRDF
StoryManager
OpenRDF
GameManager
OpenRDF
VisualisationManagerillustrations
NPL
Reasoner
AdaptiveEngine
Visualisation
������������
Reasoning Module
AnnotationModule
VisualisationModule
www.terenceproject.eu
enriched annotations
story
texttext text
text
annotations
SemiRautomated*generation
ALS LayerGUI Layer
Learner
EducatorExpert
Learner GUI
Expert GUI
Persistence Layer
OpenRDF
UserManager
OpenRDF
StoryManager
OpenRDF
GameManager
OpenRDF
VisualisationManagerillustrations
NPL
Reasoner
AdaptiveEngine
Visualisation
������������
Reasoning Module
AnnotationModule
VisualisationModule
www.terenceproject.eu
text text text
textimage
image image image
enriched annotations
story
annotations
SemiRautomated*generation
www.terenceproject.eu
text text text
textimage
image image image
enriched annotations
story
annotations
SemiRautomated*generation
www.terenceproject.eu
template visual
text text text
textimage
image image image
enriched annotations
story
annotations
SemiRautomated*generation
ALS LayerGUI Layer
Learner
EducatorExpert
Learner GUI
Expert GUI
Persistence Layer
OpenRDF
UserManager
OpenRDF
StoryManager
OpenRDF
GameManager
OpenRDF
VisualisationManagerillustrations
NPL
Reasoner
AdaptiveEngine
Visualisation
������������
Reasoning Module
AnnotationModule
VisualisationModule
www.terenceproject.eu
games
template visual
text text text
textimage
image image image
enriched annotations
story
annotations
SemiRautomated*generation
ALS LayerGUI Layer
Learner
EducatorExpert
Learner GUI
Expert GUI
Persistence Layer
OpenRDF
UserManager
OpenRDF
StoryManager
OpenRDF
GameManager
OpenRDF
VisualisationManagerillustrations
NPL
Reasoner
AdaptiveEngine
Visualisation
������������
Reasoning Module
AnnotationModule
VisualisationModule
www.terenceproject.eu
text
story
text + visual
games
AUTOM. ➤MANUAL ➤ AUTOM.
SemiRautomated*generation
www.terenceproject.eu
Requirements Prototypes Analytic + small ev. Int. prod. Large ev. Fin. prod.
HOW%DID%WE%GET%TO%THE%TERENCE%SOLUTION
www.terenceproject.eu
APPROACH WITH%WHOM EXAMPLE%METHODS WHEN
analytical HMI%experts%or%domain%experts
heuristic%evaluationformative,%summativeexpert%evaluation
cognitive%walk`through
small`scale learnersobservations
formativethink%aloud
large`scale learners aield%studies summative
EVALUATION%IN%TERENCE
www.terenceproject.eu
APPROACH WITH%WHOM EXAMPLE%METHODS WHEN
analytical HMI%experts%or%domain%experts
heuristic%evaluationformative,%summativeexpert%evaluation
cognitive%walk`through
small`scale learnersobservations
formativethink%aloud
large`scale learners aield%studies summative
EVALUATION%IN%TERENCE
From D7.4
LARGE`SCALE%EVALUATION%DESIGN
Common*design*of*the*intervention*with*TERENCE:%
‣how:*pretest/posttest%design,%with%experimental%and%control%groups%‣hypothesis:%TERENCE%improves%reading%comprehension%measured%with%standardised%text%comprehension%tests
ControlExperimental
From D7.4
LARGE`SCALE%EVALUATION%DESIGN
Common*design*of*the*intervention*with*TERENCE:%
‣how:*pretest/posttest%design,%with%experimental%and%control%groups%‣hypothesis:%TERENCE%improves%reading%comprehension%measured%with%standardised%text%comprehension%tests
ControlExperimental
Control*vs*experimental%
‣ analysis:%repeated%measure%ANOVA%%‣ results:*the%hypothesis%is%conairmed%(F=25.65,%p<0.0001)%
From D7.4
Pre`post%performances%for%text%comprehension%(dependent%variable)%were%as%follows:%‣Pescina:%` pre:%14%poor%comprehenders%(20.59%)%%` post:%6%poor%comprehenders%(8.82%)%
‣Avezzano:%` pre:%15%poor%comprehenders%(5.95%)%%` post:%2%poor%comprehenders%(0.79%)
EXPERIMENTAL%GROUP
PrePescina Avezzano
5,95%
20,59%
From D7.4
Pre`post%performances%for%text%comprehension%(dependent%variable)%were%as%follows:%‣Pescina:%` pre:%14%poor%comprehenders%(20.59%)%%` post:%6%poor%comprehenders%(8.82%)%
‣Avezzano:%` pre:%15%poor%comprehenders%(5.95%)%%` post:%2%poor%comprehenders%(0.79%)
EXPERIMENTAL%GROUP
PostPescina Avezzano
0,79%
8,82%
From D7.4
Pre`post%performances%for%text%comprehension%(dependent%variable)%were%as%follows:%‣Pescina:%` pre:%14%poor%comprehenders%(20.59%)%%` post:%6%poor%comprehenders%(8.82%)%
‣Avezzano:%` pre:%15%poor%comprehenders%(5.95%)%%` post:%2%poor%comprehenders%(0.79%)
EXPERIMENTAL%GROUP
PostPescina Avezzano
0,79%
8,82%
‣Wilcoxon%signed`rank%test%supports%that%differences%are%statistically*signi6icant*` Pescina:%z=`4.904,%p<0.0001%` Avezzano:%z=`2.266,%p=0.0234From D7.4
EXAMPLE%METHODS%IN%TERENCE
APPROACH WITH%WHOM EXAMPLE%METHODS HOW
analytical HMI%experts%or%domain%experts
heuristic%evaluationformative,%summativeexpert%evaluation
cognitive%walk`through
small`scale learnersobservations
formativethink%aloud
large`scale learners aield%studies summative
From D4.2 and D4.3 technical annex
From D4.2 and D4.3 technical annex
overall%generation
text
story
text$+$visual
games
revision%of%Natural%Language%Processing%(NLP)%generation%of%texts%for%labelling%%%%%%%%%
text
text text text
Evaluation*tasks
revision%of%Automated%Reasoning%selection%of%central%events%and%solutions
From D4.2 and D4.3 technical annex
overall%generation
text
story
text$+$visual
games
revision%of%Natural%Language%Processing%(NLP)%generation%of%texts%for%labelling%%%%%%%%%
text
text text text
Evaluation*tasks
revision%of%Automated%Reasoning%selection%of%central%events%and%solutions
Automated*reasoning*selection*of*central*events*for*games:%%>Results:%only%in%15%out%of%250%cases%(6%),%it%was%necessary%to%select% a% different% central% event% than% the% automatically%generated%one
From D4.2 and D4.3 technical annex
Selection*of*central*events
Automated*reasoning*selection*of*central*events*for*games:%%>Results:%only%in%15%out%of%250%cases%(6%),%it%was%necessary%to%select% a% different% central% event% than% the% automatically%generated%one
From D4.2 and D4.3 technical annex
>Implications2 for2 AR:%none%picked%up
Selection*of*central*events
Automated*reasoning*selection*of*plausible*solutions:%%>Results:%out%of%140%changes%of%selection%of%solutions,%the%majority%was%for%wrong%solutions
Selection*of*solutions*and*distractors
From D4.2 and D4.3 technical annex
Automated*reasoning*selection*of*plausible*solutions:%%>Results:%out%of%140%changes%of%selection%of%solutions,%the%majority%was%for%wrong%solutions>Implications2 for2WP4:*new%heuristics% for%wrong*plausible*solutions%in%the%last%part%of%Y3,
Selection*of*solutions*and*distractors
From D4.2 and D4.3 technical annex
Automated*reasoning*selection*of*plausible*solutions:%%>Results:%out%of%140%changes%of%selection%of%solutions,%the%majority%was%for%wrong%solutions
` generate% a% wrong% solution% from%correct% one%by% changing%participants,%e.g.,%<correct_sentence%id="2">The%man%ran%and%fell%on%the%ground.%%%%%</correct_sentence>%<wrong_sentence%id="2wh1">Peter%ran%and%fell%on%the%ground.%</wrong_sentence>%
>Implications2 for2WP4:*new%heuristics% for%wrong*plausible*solutions%in%the%last%part%of%Y3,
Selection*of*solutions*and*distractors
From D4.2 and D4.3 technical annex
Overall*generation:%development%times:%%>Results2for2revision2time:22` 12’6”%per%game%instance:%↑%12’8”%for%time%games%↓%10’6”%for%who%games%>Results2for2creation2time:22` avg.%23”%per%game%instance
text
story
text$+$visual
games
From D4.2 and D4.3 technical annex
Overall*generation
Overall*generation:%development%times:%%>Results2for2revision2time:22` 12’6”%per%game%instance:%↑%12’8”%for%time%games%↓%10’6”%for%who%games%>Results2for2creation2time:22` avg.%23”%per%game%instance
text
story
text$+$visual
games
From D4.2 and D4.3 technical annex
>Implications2for2the2generation2process:%the%AI`based%semi`automated%development%process%seems%to%be%promising%for%optimising%development%times
Overall*generation
OUTLINE
1. Student%centred%learning%and%technology%enhanced%learning%
2. The%TERENCE%case%study%
3. Game%over%
AND%THE%QUIZ%SOLUTION%IS…
MAIN%REFERENCES1. R.%Gennari,%S.%Tonelli,%and%P.%Vittorini.%An*AIRbased*Process*for*Generating*Games*from*Flat*Stories.%In%Proc.%of%the%33rd%SGAI%Conference%(AI`2013),%Cambridge,%UK,%December%2013.%
2. F.%de%la%Prieta,%T.%Di%Mascio,%R.%Gennari,%I.%Marenzi,%P.%Vittorini.%User*Centred*and*Evidence*Based*Design:*an*Experience*Report.%In%IJTEL%Journal,%Inderscience%Enterprise%Ltd.,%forthcoming%in%2014.%
3. D.%Di%Giacomo,%R.%Cecilia,%V.%Cofini,%T.%Di%Mascio,%R.%Gennari,%P.%Vittorini.%Silent*Reading*Supported*by*Adaptive*Learning*Technologies:*
Influence*In*the*Children*Outcomes.%In%Computers%in%Human%Behavior%Journal,%Elsevier,%forthcoming%in%2015.%
4. TERENCE%deliverables%for%technical*annex*to*D4.3.1,*D4.3.2,*D4.3.3%and%D7.4:%http://www.terenceproject.eu/web/guest/public`deliverables%%
Merry Seasonal Festivitieswith
TERENCE booksfor remembering
its best stories
THE TERENCE BOOK 2THE TERENCE BOOK 1
Dedicated to:TERENCE team, reviewers, teachers, children, and all those that contributed to the TERENCE stories. Our sincere “grazie”.