a journal of a researcher

Tuesday, February 07, 2006


Modeling intelligence

What we did with monotonicity analysis, for me, is knowledge abstraction. Different from learning, which is hard to link to existing knowledge, model abstraction has to link to domain knowledge and more important, it has to show its correctness in problem solving. Different from learning, it is to build model from numerical data. Different from parameter recognition, it has to explain the principle of the physical system. I hope it can open a new era for knowledge representation, discovery and modeling. This brings me to machine learning and data mining.

Formal modeling and model based reasoning

I think I will build my specialty in modeling using formal methods, and modeling with model abstraction, modeling vs. problem solving. I need to improve my skill in formalization and algorithm design.

Distributed intelligence

Intelligent computing in distributed systems, including distributed diagnosis


Post a Comment

<< Home