EAISI lecture by visiting Professor Mark Coates

Mark Coates, Professor in the Department of Electrical and Computer Engineering at McGill University of Montreal, is a guest of Rui Pires da Silva Castro, Associate Professor at the Statistics Group of the department of Mathematics & Computer Science, TU/e.

Title  |  Learning on Graphs: Advances and Applications 

In numerous settings, ranging from medical diagnosis to quantitative finance, we observe interacting entities and need to make predictions based on the observed relationships. In many cases, we can represent such data using a graph, with nodes representing the entities and the edges depicting the relationships and interactions. Over recent years, there have been major advances in techniques for representing and learning from graph data. In this talk, we will discuss some of the important new techniques, including graph transformers and enhanced graph convolution. We will highlight both the strengths and weaknesses of the techniques, focusing on representational power and computational overhead.  It is important to develop methods that are robust to graph errors such as missing or spurious edges, and it is highly desirable that an inference technique can provide confidence bounds. We will introduce a Bayesian graph learning framework that delivers the desired robustness and uncertainty characterization. We will conclude by highlighting some of the practical applications of the graph learning techniques, including recommender systems and circuit design.

Program
15:30 - 16.15   Lecture in Neuron 0.262 (doors open at 15:15)
16:15 - 16:30   Q&A
16:30                  Drinks at EAISI office, 1st floor Neuron