Date:  3rd May.

Title：Logic Deduction and Machine Learning in a Uniform Framework

Abstract：

Association analyses aim to disclose inherent relations and regularities between entities. No matter how desirable, the study is quite challenging, especially when dealing with graph-structured data. Is it possible to unify logic rules and machine learning to capture hidden associations and better still, explain the discovered associations? Would it make association analyses harder if we put logic deduction and machine learning together? What practical value can we get out of it after all? How can we efficiently discover useful associations from a big graph with billions of vertices and trillions of links? Moreover, can we still practice association analyses when we do not have high-quality graphs in place? This talk tackles these issues.