Graph Learning On Wednesdays

GLOW is a new reading group designed to foster discussions on the foundations and latest developments in Graph Machine Learning.

We meet (once a month) on Wednesdays at 5 pm CET

Join us on Zoom

Join our mailing list

Join our calendar

Join our Slack

Upcoming Event

26th March 2025

LLMs as GNNs (to understand how they generalize)

Problems such as efficient (length) generalization are highly relevant for building agentic systems, yet they have puzzled LLM researchers for several years. However, they can sometimes be obvious through the lens of graph representation learning. This is because LLMs are, by their design, vulnerable to generalization issues from a variety of geometric perspectives.

A typical GNN researcher will already likely have a solid understanding of most of these perspectives! In this talk, we will use this knowledge, and attempt to scale the mountain of LLM generalization ⛰️

Petar Veličković

Research Scientist

Google Deepmind

Our Concept

GLOW aims to create an inclusive and accessible environment that encourages interaction, especially for junior researchers and those outside major research labs.

Interactive Discussions: Unlike traditional reading groups dominated by long presentations, GLOW emphasizes small group discussions, breakout rooms, and active participation.
Diverse Formats: Our sessions will rotate through various formats, including 30-minute paper presentations followed by discussions, expert panel deep dives on specific topics, interview-style Q&A sessions with authors, and spotlight sessions for spontaneous brainstorming or early-stage ideas.
Thematic Focus: We plan to explore big themes over multiple sessions, such as invariance, expressivity, and the role of transformers in graph learning.