Foundational Work
- We also did some foundational work along with CABot.
- I implemented a variable binding mechanism using short term
potentiation. Most VB work is done by synchrony.
- I've been doing some work on Symbol Grounding with Fawad
Jamshed.
- A variant of CABot1 was used to learn several 2-D shapes, and
label them.
- It managed symbolic theft.
- It also learned context sensitive categorising, a type of functional
grounding.
- Hina Ghalib is about to defend her thesis on sequence learning with
CAs. This takes advantage of sequence CAs, that are dynamically
learned.
- There are a range of associative memory systems that Kailash has
developed. These are many to many, context sensitive, and involve
overlapping CAs.
