Research In Multi-Agent Learning Algorithms

点击翻译按钮获取中文摘要

The research project involves the following three levels: 1) work in theoretical inductive inference towards the discovery of general properties of learning algorithms which can indicate whether a particular problem class is learnable or not, and the precise relationships between the various components involved in the learning paradigm; 2) work in the development of feasible learning algorithms towards the discovery of general classes of problems which can be feasible learned (i.e., learned in polynomial time); and 3) work in the application of analytical tools to specific learning systems. The majority of effort will be spent on the last two levels where the investigation into feasibly solvable learning problems with be guided by potential applications.