Financial Behaviour Learning and Inference Via Graph Learning for Improved Customer Care

Data driven personalised financial service is a key to improve not only customer care but also enhance social equality and stability. It can widen accessibility of financial products, lower cost of service, and create innovative but responsible products. However, understanding individual financial behaviour and their interactions and inter-dependencies are not an easy
task. In this work we focus on graph deep learning as the means to achieve the objectives.

Leveraging graph neural networks has provided jconsiderable benefits and advances in the fields of search and recommender systems. These solutions excel in settings where the underlying graphs and data are static, while handling evolving, and dynamic graphs remains an open problem. The latter setting is, however, crucial towards achieving high quality,
real-time predictions.