Artificial Intelligence architecture, written by François Parmentier.
Acronym for Blackboard, Agents, Concepts, Examples, Temperature.

It is based on Mitchell and Hofstadter's Copycat.
It is a creative architecture.
It is an hybrid between multi-agents systems, neural networks, and semantic networks (yes, it is a mix of symbolism and connectionism).
It is written in C++.

BAsCET, like Copycat, is better than other classical systems: although it produces only sub-optimal solutions, it is more tolerant, and behaves well in front of an unknown-typed problem.
Its determinism management (through the temperature mechanism) allows it to begin with envisioning many kinds of solution, and, with the improvement of the current solution, to make more and more deterministic and relevant assumptions.
Its knowledge comes from general models and also from very specialized procedural knowledge.
Its functioning matches human reasoning, in the sense that ideas association allows it to go from one concept to another (when they are conceptually near).

It has already been applied to the recognition of bibliographic references, and on The Traveling Salesman Problem.

Here is a list of nodes concerning BAsCET, it is a kind of table of contents for it:

Log in or register to write something here or to contact authors.