Current Problems

There are a series of problems I'm currently working on solving.
Richard Bowles and I will continue with the categorisation and associative memory work.
Sequences
1. We'd like to get some sequences working.
2. This would help us move beyond the single stable state.
3. I've got a new PhD student, Hina Ghalib, working on this.
4. We'd like to be able to recognise sequences and even complete some.
5. In addition to this computational work, we are hoping to use sequences as a basis for robotic motation.
Variable Binding
1. The classic problem is the green square and the red circle.
2. All 4 base CAs are active, but how can you tell the colour of the circle?
3. Somehow they have to be bound together.
4. Variable binding also gives us rules, which makes the whole system largely Turing complete.
5. I'd like to use binding areas of medium term memory, that can support long-term memory formation, and spontaneous activation can wipe out the binding node in the binding area.
The Grand Challenge Mind and Brain Architecture
1. The challenge is being led by Aaron Sloman, and it seems the goal will be a robot with the abilities of a 3-year old.
2. We feel that the best way to do this is to start with biologically viable neurons.
3. This involves vision, and a much better understanding of areas.
4. We're going to start with a virtual robot, and see if we can get it to do sensical things, then try a real robot.
5. We hope that this neural programming will give us a better blackboard and enable us to solve these problems coherently.