Chris' Biological Neural Nets
- Some of the work I do is in machine learning. I have a model
of biological neurons called FLIF neurons.
- They learn by a form of learning called Hebbian learning.
- We've got some work using compensatory Hebbian learning and
spontaneously firing FLIF neurons that categorises different
types of data.
- My earlier work used these FLIF neurons to build an agent
in a virtual environment.
- That did some learning, but hopefully we'll be able to
integrate this new form of learning into the agent so
that it can be much more effective.
- This all fits in with
Cell Assemblies.
- In particular, we're now looking at a more mathematical
basis of this learning mechanism, and how it would work
on a wider range of domains.
