Implementation and Evaluation of a Gradient Descent-Trained Defensible Blackboard Architecture System
This work addresses the need for trainable and defensible AI systems in domains combining neural networks and expert systems, but it appears incremental as it builds on prior proposals for gradient descent in rule-fact systems.
The paper tackles the problem of training Blackboard Architecture systems by incorporating gradient descent and activation functions, resulting in a new best path-based training algorithm that is implemented and evaluated.
A variety of forms of artificial intelligence systems have been developed. Two well-known techniques are neural networks and rule-fact expert systems. The former can be trained from presented data while the latter is typically developed by human domain experts. A combined implementation that uses gradient descent to train a rule-fact expert system has been previously proposed. A related system type, the Blackboard Architecture, adds an actualization capability to expert systems. This paper proposes and evaluates the incorporation of a defensible-style gradient descent training capability into the Blackboard Architecture. It also introduces the use of activation functions for defensible artificial intelligence systems and implements and evaluates a new best path-based training algorithm.