Ideas for Collaboration

Here are some ideas I've had.
Of course, everyone is welcome to become a member of the project team and start programming the agent.
A particularly relevant area is vision. We've got a working model of retina, V1 and V2 that recognizes lines, arcs, edges and triangles. At least Xiaohong, Carl Evans, Peter and Shahedur have expertise here that could help. However, I'm sure I've missed some and others are welcome to contribute.
Another relevant area is the virtual environment technology we're using. We're using the half-life video game environment, and people who want to stay away from neurons could help on this. We could also come up with a system to work with the BSc in Games students.
Cognitive Architecture: the long-term goal of this work is to get a full-fledged cognitive architecture running. Collaboration here could be on the architecture structure, or on getting some simulations running to account for psychological data. An architecture that helps with, say, HCI design problems is probably beyond the scope of this grant, but we could start.
Learning: a lot of prior work has been on getting the system to learn categories (categorisation). Continuing this work, and extending it to rule learning is one of the more lofty goals of this project.
Planning: we've proposed that the system follows goals based on a neural version of a spreading activation net (Maes). This has to be implemented, but other planning mechanisms could help.
Language Generation: the agent only needs to accept commands. A great extension would be for the agent to generate language to provide more interaction with the user. This could lead to a whole range of interesting interaction including
Agent collaboration: this will be a video-game agent. How would groups of these agents interact.
Computational Theory: I've thought about issues of CA formation, loss, and fractionation. If someone is interested in figuring out the underlying mechanics of this, it could really be something interesting for the neural modelling community in general.
Neurobiology: the project team knows a bit about the neurobiology of vision, language and action, but not enough.
Finally, it's entirely viable to start work on the next grant, whatever that is, now.
There is also scope for using the project as a case study in say software design, project management, HCI or research methodology.