Data Fusion [of Everything [for Everyone]]
In everyday life, people prefer to make decisions by considering all the available information, and often find that inclusion of even seemingly circumstantial evidence provides an advantage. We thought it would be great if a computational method could in a similar manner infer knowledge from large sets of coarsely related data. I will overview a recently proposed approach for large-scale data fusion that leaves all the data in its original domain space and requires little or no data engineering of the input. It assembles the data in a mosaic called data fusion graph and uses compression and chaining through the compressed system to yield a model wherein every piece of evidence counts, even if it is only distantly related to the prediction task. In the second part of the talk, I will introduce a visual programming tool that could be used to support data fusion and other data science tasks, and allow lay users to intuitively use otherwise computationally complex data science approaches.
About the lecturer
Blaž Zupan is a professor at the Faculty of Computer and Information Science, University of Ljubljana, where he heads the Laboratory for Bioinformatics. He is also a visiting professor at Baylor College of Medicine in Houston, USA. His main research interests are data mining, machine learning, data fusion, and interactive data visualization. His lab develops Orange (http://orange.biolab.si), a popular open source data mining suite. Using visual programming, users of Orange can combine basic and advanced data science components and build powerful workflows for data analytics.