Globality vs Locality: Capturing Information at Different Scales in Retail Graph Networks скачать в хорошем качестве

Globality vs Locality: Capturing Information at Different Scales in Retail Graph Networks

A talk by Gergely (Greg) Kantor, Research Data Scientist @ dunnhumby. In today’s dynamic world of retail data science, uncovering patterns in transactional data is key to understanding customer behaviour. Graph Neural Networks (GNNs) are emerging as a game-changing approach—offering a fresh lens to model the intricate web of relationships hidden in this data. But while the potential is immense, capturing rich, meaningful information at scale remains a major hurdle. In this talk, we’ll take you on a journey through how GNNs interpret transactional data differently from traditional methods—revealing a compelling interplay between local patterns and global structures within graph networks. We’ll explore how these multi-scale insights can enrich our models, helping us better reflect the underlying complexity of retail behaviour—and perhaps even reach beyond it. The Data Science Festival is the place for data-driven people to come together, share cutting-edge ideas, and solve real-world problems. We run monthly events, meet-ups, and the biggest free-to-attend data festivals in the UK. Join the community at https://datasciencefestival.com/