MI Banner

The BCS Machine Intelligence Prize is awarded for a live demonstration of 'Progress Towards Machine Intelligence'.

Summary of Finalists - 2007

FearNot! - An Interactive Anti-Bullying Software for 8-12 year olds

Mei Yii Lim and Michael Kriegel, Heriot-Watt University (myl@macs.hw.ac.uk, michael@macs.hc.ac.uk)

How can innovative graphical and artificial intelligence technologies be applied to combating the serious problem of bullying in schools? FearNot! is a virtual drama application produced by the EU-funded eCIRCUS project (www.e-circus.org) led by Heriot-Watt university. Children aged 9-11 are asked to act as an 'invisible friend' to an intelligent graphical character who is suffering bullying. Characters are not pre-scripted but choose their own actions in real-time. FearNot! was an entry in last year's competition, since when it has been improved by the addition of speech synthesis and autobiographical memory. User studies are currently underway in schools in the UK and Germany.

Speech Input Logon (SiLog)

Tony Allen and Sergio Grau, Nottingham Trent University (tony.allen@ntu.ac.uk)

In the Biometric PC logon demonstration the conventional Windows logon process is extended to include a voice verification stage. A user enters their ID and password, via the keyboard as normal, but then the Biometric logon system makes a VoIP call to a VXML server which extracts the voice characteristics of the user from a spoken dialogue. If the captured voice characteristics match those of a previously registered voice profile then PC logon proceeds as normal. If no match is possible then a potential un-authorised system access has been detected and the logon process can be aborted.

The Painting Fool

Simon Colton, Maja Pantic and Michel Valstar, Imperial College, London (sgc@doc.ic.ac.uk)

The Painting Fool is an automated painter which aims to be a creative artist in its own right. In the demonstration, we will ask for a volunteer to sit in front of a video camera, and to express an emotion, for instance by smiling or frowning. Machine Vision software will tell The Painting Fool both where the features of the sitter's face are, and their emotional state. The Painting Fool will then paint a unique portrait of the sitter, using its appreciation both of the emotion being expressed and its own abilities. It will choose a suitable colour palette, art materials, abstraction level and painting style - which may involve exaggerating features - to heighten the emotion in the portrait that it produces.

eGanges (electronic Glossed adversarial nested graphical expert system)

Xenogene Gray, Grays Knowledge Engineering (www.grayske.com)

eGanges is a 5GL expert system shell with build and consult in the same minimalistic communication interface. Its nested 2D River visualisation (cf. Ishikawa quality control fishbones) is freely navigable as Major deductive conditional propositions which may also be processed, according to the Minor premises of the user's input selected from available answers to questions for each of the antecedent nodes. Epistemological heuristics allow for super-expert combinatorics with uncertainty, consistency distinct from necessity, pro tem feedback to preserve disjunctions, and interim adversarial consequents at each point in the consultation. Available non-monotonic inductive and abductive premises are node gloss data additions.

Hybrid human-machine intelligence system to generate academic research papers

Eric Atwell, Leeds University (eric@comp.leeds.ac.uk)

For RAE, Researchers must produce 4-6 journal papers in 6 years. My hybrid of human and machine intelligence can produce 60 journal papers in 6 weeks - a BIG advance in Machine Intelligence AND Human Intelligence! MI+HI are integrated at 3 levels: use of AI architecture; intelligent use of MI tools; and student learning. Individual components are not particularly novel; the advance in MI comes through the novel combination. For the demo, a volunteer from the audience will act as a student "intelligent agent", to demonstrate how AI architecture + student + MI can generate an outline journal paper.