Keynote Speakers

Day 1 – Nov 20, 2026

Speaker: Wengong Jin

Affiliation: Assistant Professor, Northeastern University

Day 1 – Nov 20, 2026

Speaker: Yizhou Sun

Affiliation: Professor, UCLA

Day 2 – Nov 21, 2026

Speaker: Rajmonda Caceres

Affiliation: Professor, UCLA

Day 2 – Nov 21, 2026

Speaker: Petar Veličković

Affiliation: Senior Staff Research Scientist, Google Deepmind & University of Cambridge

Day 3 – Nov 22, 2026

Title: When Can AI Transfer Across Power Grids?

Speaker: Anna Scaglione

Affiliation: Stephen M. Ross/Related Companies Professor, Cornell Tech

Abstract: AI models for power systems are often trained for a particular network and set of operating conditions. But power grids change, and a model that works on one system may not work on another. How can we tell when what a model has learned will transfer?

In this talk, I will explore how the structure and physics of power networks can help answer that question. I will discuss representations that capture features of a network and its response, and how comparisons between systems can help assess whether a model trained on one grid is relevant to another. Drawing on recent work on distribution systems, I will consider both the opportunities and the limits of transfer, and what further evidence is needed before transferred models can be trusted in operation.

Bio: Anna Scaglione is the Stephen M. Ross/Related Companies Professor of Electrical and Computer Engineering at Cornell Tech and Cornell Engineering. Her research brings together statistical signal processing, network science, decentralized optimization, and machine learning, with applications to communication networks and energy infrastructure. She directs the Signal, Information, Networks and Energy laboratory, whose work includes graph learning and foundation models for the electric grid. She is an IEEE Fellow.