Final Report

IST-2001-33310 VICTEC

January 2005

Final Report

AUTHORS: R.S.Aylett

STATUS:Draft

CHECKERS:S.A.Richardson/S Louchart

Project Manager

Name: Professor Ruth Aylett

Address: CVE, Business House, University of Salford, University Road,, Salford, M5 4WT

Phone Number: +44 161 295 2922

Fax Number:+44 161 295 2925

E-mail:

1- Summary of the state of the project and objectives attained

VICTEC was an ambitious project. Firstly it sought to apply novel technology to a branch of education – Personal and Social Education in which little technology of any type had been applied. Within this area, it picked education against bullying, a difficult area since there is no magic-wand solution to bullying behaviour and considerable variation as a result in conventional pedagogical objectives. Secondly, it sought to further innovate the technology itself, developing the concepts of Empathic Characters and Emergent Narrative. Thirdly, it sought to evaluate its technology on a larger scale than any previous project in synthetic characters, with 750-800 children involved in its formal evaluation exercises apart from all those involved earlier in piloting questionnaires and user-centred design of characters and stories.

Necessarily, VICTEC was also an inter-disciplinary project. In computing, it drew on artificial intelligence, 3D interactive graphics and HCI experts. In psychology, it involved field psychologists with experience of investigating bullying as well as theoretical psychologists investigating architectures for personality and affect. It also drew on educational expertise both from project members and all the associated teachers. Thus alongside its scientific, technological and educational challenges, it had to find ways of reconciling different professional cultures and objectives.

Finally. VICTEC had to meet the day-to-day trials of any human endeavour involving a number of people. Real babies were not a project objective, but six were born to project members over the 29 months, each with their own package of maternity leave and sleepless nights. Researchers got jobs elsewhere, one of the partners had a change of corporate objectives. In this report we show how the project met all these challenges and achieved its objectives and reflect on what the team have learned as a result of working together.

The project brought together five partners. University of Salford, in the UK, coordinated the project and contributed to the formulation of a theory of emergent narrative, development of the authoring tool, the development of characters, a language system and some of the evaluation activity. University of Hertfordshire, the other UK partner, with many years experience of fieldwork on bullying, led on work with schools, the development of scenarios, pedagogical evaluation methodology, inclusing the bullying questionnaire, and with Lynn Hall, of Sunderland University, user-centred design with children and usability evaluation.

As theoretical psychologists, Bamberg Technical University, in Germany, contributed theoretically to the project’s understanding of empathy and personality, designed the psychological evaluations and empathy questionnaire and assessed aspects of empathy in relation to synthetic characters experimentally. INESC-ID, in Portugal, developed the synthetic character architecture, the overall Framework for virtual dramas with synthetic characters and the FearNot! demonstrator, as well as carrying out the pedagogical evaluation in Portugal. Until a change in corporate policy reduced their participation, ONI-AUTOR, a Portuguese company, contributed to graphical character design, multi-media aspects of the FearNot! trailer, and to ideas for exploitation of project outputs.

In summary the project’s key achievements are:

•The production of The FearNot! demonstrator applying virtual drama to education against bullying with versions in English, Portuguese and German

•The large-scale evaluation of FearNot! in the UK, Germany and Portugal with a total sample of around 800 children

•The involvement of schools in all three countries in the design of characters and scenarios for FearNot!

•The development of a generic Framework uniting AI and graphics technologies for the production of virtual dramas with synthetic characters

•The production of a toolkit for experts to configure FearNot! scenarios and characters

•The development of a theoretical basis for the empathic relationship between human users and synthetic characters, and the specific role of ideomotoric empathy

•The development and evaluation of an intelligent toy as an interface to FearNot!

• Empirical investigation of the relationship between the theory of mind and bullying roles and between empathic capacities and bullying roles leading to new results.

•The creation of a network of involved schools in all three countries and the building of links between teachers in these schools and the project, including a successful teachers’ workshop

•Very wide dissemination of the project’s work through presentations, conference and journal papers, six technical workshops, media reports, a pamphlet, videos, and project brochure

2 – Project Objectives

In line with its multi-disciplinary nature, VICTEC had objectives in a number of different areas: pedagogical; scientific; technology development and technology production. These objectives are reproduced from the Technical Annex:

Pedagogical Goals and Challenges

P1: Allow children to explore perspective taking and reflection for dealing with complex social problems (WP5; WP6).

P2:Engage teachers and educational science researchers in the design process for such environments (WP2)

P3: Assess the impact that such environments will have on children's social relations. (WP7)

But VICTEC also had a larger and broader pedagogical impact:

P4 Demonstrate to the education and training community technology for educational applications in personal and social education using a new kind of technology (VEs) and in particular virtual improvisational drama. (WP6)

P5: Going beyond the violence and bullying problems, VICTEC also investigated the potential of Virtual Environments (VE) to enhance the school curriculum across Europe.

P6: VICTEC developed a network and representative panel of users to advise on development of the project and share the progress and outcomes of the project raising awareness of the potential for using VE in schools. (WP1)

Scientific Goals and Challenges

Rather than trying to build yet another synthetic character, VICTEC created a framework for the creation and analysis of VEs inhabited by synthetic characters. It developed certain properties of synthetic characters, focusing on their empathy, believability and credibility. This framework was used as the basis for the development the FearNot! interactive drama application, evaluated in schools across Europe. The project’s scientific objectives were:

S1. To conceptualise the process of perspective-taking: seeing the (social) world through the eyes of somebody else, re-experiencing situations from different viewpoints, i.e. adopting different social roles. (Task 3, WP5)

S2. To develop a theory of emergent narrative from which computational requirements could be derived. (Task 1, WP2; Task 3, WP3)

S3. To establish a theory of social agent personality (Task 4, WP5)

S4. To analyze available evaluation techniques and map their characteristics onto the evaluation requirements of synthetic characters and their empathic relations with children (Task 5, WP2)

S5. To establish the criteria for the use of an intelligent toy as an interface (Task 2, WP2)

S6. To disseminate our results and tools throughout the European schools and the world, aiming to be one of the main driving forces in the advances that will be taking place world-wide in this area of research (WP8).

Technology development

If research challenges address the conceptual framework, the technological challenges address issues related to the creation of a toolkit including:

T1: To design a general framework for empathic synthetic autonomous agent creation covering cognition, behaviour and embodiment (Tasks 1&2, WP3)

T2: To incorporate external physical devices (intelligent toys) supplied with sensors as user interface devices (Tasks1-4, WP4)

T3: To establish a set of design guidelines for believable and empathic synthetic agents (Tasks1-4, WP5)

T4: To produce a toolkit consistent with both agent and graphical standards (Tasks 3-5, WP3)

Technology production

VICTEC constructed the FearNot! demonstrator which was translated into three languages: English, Portuguese and German. This demonstrator had a multifaceted role in the project:

• To prove experimentally the validity of the conducted research concerning the development of synthetic characters and the “empathic relation established” between children and such characters.

• To illustrate the use of the framework in innovative and practical examples, where empathy, believability and credibility of synthetic characters play a fundamental role

.

3 – The achievements of the project (unfinished)

3.1 Pedagogical achievements

Allowing children to explore perspective taking and reflection for dealing with complex social problems – in this case bullying - through the FearNot! demonstrator

Engaging teachers, children and educational science researchers in the design process for such environments using, Kar2ouche story-boarding software and the FearNot! trailer, as well as presentations and discussions at for a like the April 2003 Childline conference and the Teachers’ workshop organised in Lisbon in October 2002. Involvement in the Teachers’ workshop held by the Minerva project VES in Kassel in October 2003.

Large-scale evaluation of the impact of the technology through the June 2004 Evaluation event at University of Hertfordshire, involving 400 children, and the separate pedagogical evaluations in the UK, Portugal and Germany, involving more than 100 children in each.

In conjunction with the Minerva project VES (virtual Environments in Schools) helping to investigate the general uses of virtual environments in the school curriculum. The VES-developed 3D parachute drop was included in the programme activities of the Herts evaluation event and provided a large-scale evaluation of the use of the package for science education.

3.2 Scientific achievements

Creation of a theory of empathic synthetic characters and a supporting agent architecture

New results on what aspects of human empathy allow a relationship to be built with synthetic characters: the role of ideomotoric empathy, the role of similarity

Development of a theory of emergent narrative

New resulats on links between theory og

3.2.1 Dissemination

The project has carried out a very substantial amount of dissemination by a wide variety of channels.

Book chapter

Journal papers

Conference papers

Workshop papers

Workshops organised

EU-level events participated in

Exhibition

Press and other public media publicity

University seminars

Work with VES and ELVIS

3.3 Technological achievements

FearNot!

The Framework

4 – Methodologies (unfinished)

User-centred design

Iterative prototyping

Bottom-up testing (as against the more usual top-down testing)

The need for large-scale evaluation

Use of questionnaires: development and piloting

The use of qualitative evaluation: focus groups

The difficulties of evaluating empathic effects in short-term interaction: the need for a longitudinal study

5 – European added value (unfinished)

5.1 Links with related projects, attendance at EU-wide events

Participation in EU-level events: Concertation meetings, ProLearn workshops, the Lisbon exhibition, OnlineEduca (2003 and 2004)

Links with Minerva through VES and also to work in China through ELVIS

Magicster (Gesticon workshop) Cogniron (activities included in the Herts evaluation prgramme)

5.2 Lessons learned

Sensitivity to cultural differences over three copuntries: behaviour of children in the target age-group differs; perceptions of what is aggressive language and behaviour; what is appropriate school clothing; where bullying takes place

Differences in primary education and the school and general culture in the three countries.

ICT in education (for example availability of computers in the classroom) very different in the three countries, as is level of awareness of bullying as an issue and approaches to tackling it.

Amount of technical resource required was much more than we had realised, especially for bottom-up systems: lesson of The Sims too

The need for early prototypes in order to run early evaluation and involve end-users in design. We hadn’t planned to do this but it turned out to be essential

Content, content, content. Much more input from graphic designers was needed than had been realised. But of course this isn’t really a research objective and researchers may not be the best people to produce it. Here AUTOR could have helped more but for their change of corporate strategy

The lack of supporting graphical tools, environments and standards for synthetic character work is a major obstacle to exploitation in schools: the use of proprietary games engines can solve the problem for research but not at all for exploitation

The lack of easily-available language and speech engineers in the first case added an extra item into the WP and in the second case proved a major issue for the unscripted system

To work together, everyone has to accept the need of everyone else to get something out of the project: all researchers need to do publishable research; all companies need commercially exploitable technology.

Ability to understand different viewpoints and professional cultures and to understand some of the basic issues in the non-home discipline was developed. The tension between having early stable technology to support evaluation and technology innovation, which tends to produce late and rather flakey prototypes had to be overcome.

There is no substitute for face-to-face contact: trips to work alongside each other really do help. Keeping people communicating is really very hard when geographically dispersed and separated by language and culture. A major task of the project coordinator is keeping in close touch with everyone and using the best modality for this, whether email, chat or telephone meetings.

The issue of building links with the overall technology-enabled learning research community given that our approach seems very innovative compared to most others we have seen.

6 – Outlook (unfinished)

A short description on how the results and achievements of the project have benefited each partner and how the partners intend to use and exploit these further.

Salford: links in to other work at CVE and to work in Art and design, to project ELVIS in collaboration with Chinese group; possibility of putting FearNot! into an immersive display system; using the work as a basis for two FP6 bids in call 4. Using the Framework for PhD and other student projects Links into the FP6 NoE Humaine. Publications.

INESC-ID: opensource development of the Framework; use of the character architecture in further projects; using the work as a basis for two FP6 bids in call 4. Using the Framework for PhD and other student projects Links into the FP6 NoE Humaine. Publications.

Herts: Links into the other work the group with robots in areas of social interaction with children (FP6 Cogniron project), including work with autistic children; using the work as a basis for two FP6 bids in call 4. Contributes to the continuing research on possible bullying intervention strategies. Benefits our links and working partnership with schools. Publications.

Bamberg: Links into further work on empathy especially ideomotoric empathy and to other projects. Links into continuing work on the PSI model of personality, affect and action-selection. Using the work as a basis for two FP6 bids in call 4. Publications.

AUTOR: originally hoped to derive exploitable software and improvements to characters on their large-scale website for children. Due to a change in corporate strategy by ONI, the current owners of AUTOR, no exploitation path is now foreseen there

We see the gap between the project innovation and the current state-of-the-art as a potential problem for commercial exploitation of the results though we have addressed this issue in our Exploitation Plan. Following discussion at the review meeting in June, it was felt that an opensource path to exploitation would be more successful, and the project has involved itself in related work funded by the Asia-Europe IT&C called ELVIS under which synthetic character facilities for the open source rendering engine OpenSG are being developed (agent-SG – see The generic run-time framework underlying the FearNot! demonstrator seems a strong candidate for opensource exploitation and the sourceforge site already being used for the agent-SG project will be used to make some of the project outputs publicly available.

The team would like to secure funding to continue both to develop the whole concept of empathic agents and to explore their wider application in other areas of personal and social education. The project is discussing initiating two proposals under call 4 of FP6: one in cognitive systems called EMINENTS (EMotional INtelligence for Embodied ageNTS) and one in Technology-Enhanced Education called eCIRCUS: Education through Characters with emotional Intelligence and Role-playing Capabilities that Understand Social interaction.

7 – Conclusions (unfinished)

Annex 1

Deliverables

References

Annex 2

The VICTEC brochure (see separate document)

- 1 -