On January 11, 2017, the 4th SFB 901 seminar in the winter semester will take place
16:00 - 17:45
Title: Non-functional analysis of software services: A machine learning approach
Speaker: Prof. Dr. Eyke Hüllermeier + Vitalik Melnikov (TP B3)
Abstract:
One of the goals of project B3 for the second phase of the CRC is the use of machine learning methodology as an alternative to the hitherto applied simulation-based techniques for the non-functional analysis of service compositions. In this seminar, we report on first steps we made in this direction and the results achieved so far. The first part of the seminar will provide some basic background from machine learning, mainly to position the problem from a machine learning perspective and to highlight the challenges it comes with. In the second part, we present the learning-to-aggregate approach that emerged within the course of this project. Moreover, the connection to our idea of OTF machine learning will be explained, and some preliminary results will be presented.