the terence fp7 project s&t goals and results for schools and beyond
DESCRIPTION
The presentation of how the TERENCE FP7 EU project met its scientific and technological goals, and worked for its beneficiaries, e.g., schools, at the final review meeting of TERENCETRANSCRIPT
Rosella Gennari!!KRDB - CS Faculty - UniBZ!!
http://www.terenceproject.eu
TERENCE !its S&T goals and potential impact for beneficiaries
© R. Gennari, LUB, <[email protected]> TERENCE 3rd review meeting, 27th Nov 2013, l’Aquila
Outlinetwo stories from the DoW book
Chapter 1 What TERENCE has done for its 4 S&T goals
Chapter 2 What TERENCE has done for its beneficiaries and its potential impact
© R. Gennari, LUB, <[email protected]> TERENCE 3rd review meeting, 27th Nov 2013, l’Aquila
characteristation of learnersFirst of all, TERENCE will provide a refined characterisation of poor comprehenders in regard to the TERENCE stories and reasoning games.
The TERENCE S&T1 goal
WP1:%user%requirements
WP7:%evalua3on
© R. Gennari, LUB, <[email protected]> TERENCE 3rd review meeting, 27th Nov 2013, l’Aquila
The TERENCE S&T1 goal
It will analyse the domain knowledge of (causal-)temporal representation, reasoning and visualisation for and with poor comprehenders.
User requirements will be essential in producing (and classifying) adequate stories and games, and in structuring the conceptual model of the ALS.
characteristation of learners
WP1:%user%requirements
WP7:%evalua3on
WP2:%story,%game,%GUI%and%ALS%design
WP6:%GUI%and%ALS%design
The requirements will thus contribute t o s h a p e t h e guidance mechanism o f the TERENCE ALS.
S&T1 will require a strong cross-disciplinary effort, starting from the available domain knowledge, and building on research results from medicine, cognitive and educational psychology.
© R. Gennari, LUB, <[email protected]> TERENCE 3rd review meeting, 27th Nov 2013, l’Aquila
The TERENCE S&T goalevaluation tasks and guidelinesTERENCE will design and assess novel evaluation tasks of text comprehension
⧂The policemen fell asleep after leaving the office ⧂The policemen fell asleep before leaving the office ⧂The policemen fell asleep when going out from the office
2
WP1:%user%requirements
WP7:%evalua3on
© R. Gennari, LUB, <[email protected]> TERENCE 3rd review meeting, 27th Nov 2013, l’Aquila
The TERENCE S&T goalevaluation tasks and guidelinesand, in particular, novel metrics for the reasoning complexity of the TERENCE games [and stories], that are common to UK and Italy.
The results of the evaluations will allow us to define novel evaluation guidelines for those in charge of evaluating educational comprehension tools for children.
metrics/indices guidelines
stories Coese D2.2 TR
games generation algos D2.3
2
WP1:%user%requirements
WP7:%evalua3on
WP2:%story,%game,%GUI%and%ALS%design
WP6:%GUI%and%ALS%design
© R. Gennari, LUB, <[email protected]> TERENCE 3rd review meeting, 27th Nov 2013, l’Aquila
The TERENCE S&T goalAI technologies and toolsTERENCE will also study scientific questions arising in automated understanding of stories for novice comprehenders, in English and in Italian.In particular, TERENCE will study and contribute to automated reasoning, possibly with explanatory feedback, that is adequate to poor comprehenders and their educators.
3
relation type
if with signal...
else if deduced...
WP1:%user%requirements
WP7:%evalua3on
WP3:%story%annota3on
WP4:%game%design%II,%txt%game%
genera3on
WP2:%game%design%I
WP4:%game%genera3on%process%design
© R. Gennari, LUB, <[email protected]> TERENCE 3rd review meeting, 27th Nov 2013, l’Aquila
The TERENCE S&T goalAI technologies and tools
The automation of TERENCE (via automated annotations and automated reasoning) will help reduce the development costs, and allow researchers to concentrate on the design of the higher-level ITS models.
As for the automation itself, TERENCE aims at improving on:
T1. novel natural language technologies and tools for story annotation, in English and Italian,
T2. temporal automated reasoning tools that can interact with the annotations of stories developed with the annotation tools,
T3. reasoning tasks and tools, so as to enhance temporal and causal-temporal knowledge, acquisition and automation of inference skills on texts.
Estimates of construction time indicate that the development of an effective ALS can be costly and time consuming.
3
learning(material(semi,automated(development
WP3:(story(annota7on
WP4:(txt(game(genera7on
WP5:(illustra7on(of(
visuals
WP6:(composi7on(of(txt(and(visuals
© R. Gennari, LUB, <[email protected]> TERENCE 3rd review meeting, 27th Nov 2013, l’Aquila
The TERENCE S&T goalmodular and reusable architecture
The difficulty in sharing components among different designers has also impeded the smooth design and development of ALSs. It is often the case that ALSs are written in many different programming languages that are incompatible. Frequently, implementers intertwine the components into one monolithic system.
Modelling the ALS conceptual models (adaptation, domain, environment, student) with ontologies, and developing the components of the adaptation engine as web services will ensure the highest level of interoperability and reusability.
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
4
WP2:%ALS%design
WP6:%ALS%development
WP3:%story%annota:on
WP6:%visualisa:on
WP4:%txt%game%genera:on
Educators will be able to adapt and reuse stories, illustrations and smart games produced by others, which will promote collaborative learning. More in general, the TERENCE web services can be reused within other collaborative web-based environments. The semantic web service architecture of TERENCE will also contribute to reducing development costs. In brief, TERENCE will contribute to the general
advancement o f ALS architectures on account of: its modularity and the reusability of its modules, developed as semantic web services.
© R. Gennari, LUB, <[email protected]> TERENCE 3rd review meeting, 27th Nov 2013, l’Aquila
Outline
Chapter 1 What TERENCE has done for its 4 S&T goals
Chapter 2 What TERENCE has done for its bene6iciaries and its potential impact
two stories from the DoW book
© R. Gennari, LUB, <[email protected]> TERENCE 3rd review meeting, 27th Nov 2013, l’Aquila
For learnerswhat we did in words
Keywords: improved reading; appetite for reading he usage of the TERENCE adaptive learning system (ALS), with adequate learning material and GUI, calibrated by an intervention plan designed by experts,
improved reading comprehension The TERENCE intervention may have also stimulated its learners' organic functioning and global cognition (e.g., visual memory, global-deductive reasoning, attention). Consequently, the longer-term impact of TERENCE may be by far larger: a potential increase in the communicative and relational competencies of its learners for communicating with peers and adults (parents and others), as well as in their self-esteem.
For instance, at the conclusion of the intervention in schools, a number of children asked to continue reading the TERENCE books at home, because they wanted to finish their stories. No better result could be hoped for the end of a project that, in its infancy, wished to nurture a desire for reading.
© R. Gennari, LUB, <[email protected]> TERENCE 3rd review meeting, 27th Nov 2013, l’Aquila
For learnerswhat we did in pics
© R. Gennari, LUB, <[email protected]> TERENCE 3rd review meeting, 27th Nov 2013, l’Aquila
For educators and parentswhat we did in wordsKeywords: novel teaching; off-the-shelf open learning material; guidelines and tools for creating adequate material Our findings also suggest the possibility of strengthening the paper- and-pencil didactic activities with the adoption of technological supports, both in an individual setting and in a social context (e.g., in the classroom). In view of that, The TERENCE ALS with its learning material and learner GUI promises to improve the quality of work of educators and families
by integrating classical teaching activities with novel learning experiences and methodologies that are flexible, since they are based on adequate stories and games that are provided in an adaptive fashion, and effective, since TERENCE improved reading comprehension.
Moreover, the use of the TERENCE authoring tool and guidelines, in particular for stories and games, obviates the need for educators to find adequate learning material, render texts adequate to poor comprehenders or deaf children, or even to manually create complex games [...] and track the progression of learners Thus, for educators, TERENCE is a time saving and cost saving tool, facilitating the adaptation of the material to learners and enabling its creation, as well as the automated generation of games out of any text.
© R. Gennari, LUB, <[email protected]> TERENCE 3rd review meeting, 27th Nov 2013, l’Aquila
For education and psychology actors
what we did in wordsKeywords: remediation; large-scale comparative analysis; evaluation tasks Potentially, TERENCE entire ALS, learning material and guidelines or authoring tool might be also used in
remediation settings where clinical psychologists may control in a non-intrusive manner (e.g., crucial for autism) the effectiveness of the remediation, and the stimulation can be also delivered at home (in a more familiar environment) for longer periods (in addition to the clinical activities).
Moreover, TERENCE did the first a large scale comparative analysis of the linguistic and cognitive profiles of hearing and deaf poor readers, in UK and Italy. It significantly improves our current understanding of the specific factors causing reading comprehension problems in the two groups of learners.
New evaluation tasks for the assessment of children's understanding of causal and temporal linguistic relations were designed and tested in two languages by large scale studies.
© R. Gennari, LUB, <[email protected]> TERENCE 3rd review meeting, 27th Nov 2013, l’Aquila
For education actorswhat we did in pics
© R. Gennari, LUB, <[email protected]> TERENCE 3rd review meeting, 27th Nov 2013, l’Aquila
For researchers and technologistswhat we did in words
Keywords: networking; scientific events; commercial exploitation As a natural consequence of evaluation results and awareness arousal, TERENCE ideas concept and vision have already been
taken up by related initiatives in Europe or senior colleagues. Examples are: the Linked Open Data (LOD) European project, in which the TERENCE learning material was published as open data, and where the annotation and reasoning modules were advertised; the MUSE European project that took up the TERENCE annotation language, methodologies and findings for NLP; the DARE research project (http://www.inf.unibz.it/dare), financed by the Province of Bozen-Bolzano, that took up TERENCE stories and game framework for co-designing games with primary-school classes. Others are reported in the TERENCE dissemination and exploitation reports.
Moreover, the TERENCE Consortium was also active in promoting TERENCE findings in research venues whether by organising events such as the evidence-based and user-centred TEL workshop series, counting already 3 successful issues, or by participating in key TEL conferences such as ICALT.
© R. Gennari, LUB, <[email protected]> TERENCE 3rd review meeting, 27th Nov 2013, l’Aquila
For researchers and technologistswhat we did in pics
© R. Gennari, LUB, <[email protected]> TERENCE 3rd review meeting, 27th Nov 2013, l’Aquila
For policy makerswhat we did in shortKeywords: awareness rising; networking with school stakeholders Last but not least, TERENCE researchers
got in touch with several policy-makers in particular in the education sector. Meetings were organised, for instance, in L'Aquila Province and with the Education Body of the
Province of Bozen-Bolzano for raising awareness of the problem of text comprehension that, in particular in Italy, is not recognised as a special development disease.Besides raising awareness of the problem, such meetings allowed consortium members to disseminate the TERENCE results and explain the novelty and potentiality of the approach of TERENCE for schools, mixing technology with traditional interventions, and visuo-perceptual skills for reasoning about stories for nowadays learners.
Support letters and agreements were signed for using TERENCE in traditional schools, some of which received ICT facilities for TERENCE.
© R. Gennari, LUB, <[email protected]> TERENCE 3rd review meeting, 27th Nov 2013, l’Aquila
S&T1characterisation
End of stories from the DoW book
WP7: evaluations
UnivAQ
WP2: story and game design
LUH
WP1: learner requirementsUniPD+UnivAQ
ucd+ebd
WP1: learner requirementsUniPD+UnivAQ
WP7: evaluations
UnivAQ
ucd+ebd
learning material semi-automated development
WP4: generation
LUB
WP3: annotation
KUL
WP5: illustration
MOME
WP6: visualisation
LUB
WP6: ALS development
USAL
learning material semi-automated development
WP4: generation
LUB
WP3: annotation
KUL
WP6: visualisation
LUB
WP2: ALS designLUH
WP7: evaluations
UnivAQ
WP2: AE rules and ALS design
LUH
WP1: learner requirementsUniPD+UnivAQ
ucd+ebd
WP6: ALS development
USAL
S&T2guidelines S&T3
AI for generation
S&T4modular !& reusable
Impact5 beneficiaries
The Endwith
TERENCE books for X-mas time
for remembering its best stories
THE TERENCE BOOK 2THE TERENCE BOOK 1
and forgetting others (sorry, Pablo!)