The BCS Machine Intelligence Prize is awarded for a live demonstration of 'Progress Towards Machine Intelligence'.
Hemin Omer Latif, Nottingham Trent University
Description of system: GazeBot is a mobile robot that demonstrates a combination of teleoperated and automated capabilities. An automated person following capability enables GazeBot to follow a moving person and keep a desired distance using one camera and based on vision cues only. While a specially developed interface called TeleGaze enables teleoperation of the GazeBot whenever is required. The TeleGaze interface enables a human operator to select the person of interest from the scene that is being monitored. Once a person being selected, the automated person following starts following and searching for the person of interest if it has lost it for any reasons.
Daden Limited, David Burden
Description of system: Halo is an automated avatar running within the Second Life virtual world. As well as being able to talk to users, navigate, and exchange objects/money, our current focus is on giving Halo emotions. Halo can learn emotional responses for objects and avatars based on their behaviour, and from talking to other users, as well as by programming. By existing (and having agency) in a rich virtual world designed for humans not computers, having a visual look identical to human users of the world, and integrating several different AI technologies, Halo provides an innovative new way to explore artificial intelligence.
Chris Huyck, Middlesex University
Description of system: CABot2 is an assistant to a user in a simulated 3D environment. It is an agent in the environment along with the user; it accepts commands from the user, views the environment, maintains its own plans and emits actions. CABot2 is entirely implemented in simulated (fatiguing Leaky Integrate and Fire) neurons. The commands are interpreted by a natural language parser that is a sound psycholinguistic model, and the initial parts of vision are a reasonable approximation of the human retina and primary visual cortex. CABot2, the second prototype in the CABot project, also learns simple actions.
Martin Rhodes and Simon Coupland, De Montfort University
Description of system: Video games are the new frontier in machine intelligence. Gamers want to play against truly human like players that make rational decisions and make mistakes in the same way a human would. Pro-Evolutionary soccer demonstrates machine intelligence through a football video game application. The current generation of football video games suffer from a predictable, repetitive style of gameplay. Pro-Evolutionary soccer uses a stochastic approach to give varied, organic style of play in the form of free-kick taking. Given a physics model and the positions of the opposing team, our system finds a near optimal free kick from all possible free kicks giving a varied and challenging gameplay.