Conclusion

• There has been a really simple point about categorisation, it's ubiquitous.
• Decision trees are also pretty simple, though learning them is more complex.
• We have talked about some really complex topics, No Free Lunch, natural kinds and biology.
• Take Home Points
  1. There is No Free Lunch
  2. Decision trees are a simple categorisation algorithm, with a question at each node, and results on the leaves.
  3. Decision trees can be learned (c4.5 is an example).
  4. Cell Assemblies are collections of neurons, that are the basis of concepts.
  5. Cell Assemblies are categorisers and can be learned.
• If you're interested in the neural modelling, talk to me. Only the simplest things will be examinable. (Integerate and fire, but not STDP.)
• Reading: For this lecture read Russell and Norvig 18.1 through 18.3 (pp 704 - 718)
• Reading: For next week read Russell and Norvig Chapter 1 (pp 66-111)