Furkan Kocayusufoglu successfully passed his PhD proposal today on "Learning with Richly Structured Data"
Committee: Ambuj Singh (Chair), Xifeng Yan, Francesco Bullo, Yu-Xiang Wang
Abstract: Learning problems often involve richly structured data, coupled with a diverse set of learning objectives depending on the application domain. The structure of the data inevitably shapes how the learning algorithms are derived for satisfactory results on the downstream task. In this talk, I will give an overview of my research about various learning tasks (e.g. prediction, generation) on spatial and temporal structures from multiple application domains such as complex networks and user modeling. More specifically, we will first discuss the benefits of incorporating temporal signals in modeling short/long-term user interests to better capture their search intent while interacting with online e-commerce and content-sharing platforms. The second part of my talk will focus on problems about a complex family of graphs called flow graphs. I will present our two recent studies aiming to (i) estimate missing flows based on graph topology and physics, and (ii) generate realistic flow graphs, providing a data-driven alternative to domain-specific simulations.