Knowledge Acquisition by Learning
Peijun Ye, Fei–Yue Wang · 2023
This chapter focuses on the acquisition of knowledge units. It constructs the fundamental “bricks” of root that the deliberative reasoning grounded. The chapter introduces some theoretical basis of learning. It then explains knowledge acquisition in an offline mode. The chapter also addresses the online learning, which extracts knowledge dynamically through behavioral records in a data stream circumstance. This adaptive maintenance can supplement new procedural knowledge to the agent's knowledge base. The chapter focuses on two methodologies – neural symbolic learning and explanation of deep learning. To test and validate the online learning method, it also applies the proposed techniques to travel behavioral analysis. Neural-symbolic learning tries to unify the symbolic reasoning and the neural network learning. With such a framework, logic and network models are studied together as integrated models of computation. This paradigm has biological foundations in cognitive science.