BICS 2004: Tutorials

Implementing neural models in silicon

Slides

Prof Leslie S. Smith, Dept. of Computing Science and Mathematics, University of Stirling Stirling FK9 4LA, Scotland: email lss@cs.stir.ac.uk

Neural models are used in both computational neuroscience and in pattern recognition. The aim of the first is understanding of real neural systems, and of the second is gaining better, possibly brain-like performance for systems being built. In both cases, the highly parallel nature of the neural system contrasts with the sequential nature of computer systems, resulting in slow and complex simulation software. More direct implementation in hardware (whether digital or analogue) holds out the promise of faster emulation both because hardware implementation is inherently faster than software, and because the operation is much more parallel. There are costs to this: modifying the system (for example to test out variants of the system) is much harder when a full application specific integrated circuit has been built. Fast emulation can permit direct incorporation of a neural model into a system, permitting realtime input and output. Appropriate selection of implementation technology can help to make interfacing the system to external devices simpler. We review the technologies involved, and discuss some example systems.

Models of Consciousness: The world scene.

Prof Igor Aleksander, Department of Electrical and Electronic Engineering, Imperial College, London

Slides

An increasing number of laboratories around the world are trying to design a machine which could be said to be conscious. Their efforts are not only revealing how to build more competent machines, they are also illuminating how consciousness arises in living beings. The key historical event was a closed conference organised by Christof Koch of CalTech and David Chalmers Arizona University in 2001. A mixture of 20 philosophers neurologists and computer scientists meeting at the Cold Spring Harbour Laboratories in the US determined almost unanimously that approaching consciousness from the perspective of computational modelling would not only introduce novel mechanisms, but would clarify many philosophical puzzles about consciousness. I review what has happened since then, drawing attention to salient work both in the US and Europe. This is proceeding over a spectrum ranging from the 'functional' to the 'material'. The functional is sited in the artificial intelligence tradition being concerned with behaviour that one would say might require consciousness, while the material is neurologically based and asks what possible mechanisms could give rise to consciousness. I amplify work in my own laboratory which is at the material end of the spectrum and breaks down into five major lines of enquiry[1]: How could a mechanism:

1. sense an out-there world with itself in it?
2. imagine either experienced or fictional worlds?
3. attend to important events in the world an in its imagination?
4. plan its future actions?
5. evaluate emotionally the nature of its plans?

1. Chalmers' 'hard' problem;
2. what it is to be unconscious;
3. animal consciousness;
4. 'illusion' theories of consciousness.

[1] I. Aleksander and B. Dunmall : Axioms and tests for the presence of consciousness in agents: Jour. Of Conc. Studies, June 2003, 15 pp.