Workshop on Learning and Reasoning with Graphs

About the workshop

Many real-world phenomena can be naturally represented as graphs, including social networks, biological systems, transaction networks, and knowledge bases. Learning and reasoning over graphs have enabled advances in applications such as drug discovery, fraud detection, recommender systems, and scientific discovery.

This workshop brings together researchers and practitioners working on graph-based AI across foundations and applications. The 3-hour session features invited talks and an interactive poster session with extended abstracts, with the goal of fostering discussion and collaboration between research and applications of these topics in the Benelux region and beyond.

Topics of interest

Graph neural networks & representation learning
Algorithmic reasoning & neuro-symbolic AI
Graph representations for relational databases
Integrations with LLMs & AI agents
Applications in science, health, finance, and beyond

Invited speakers

Jan-Christoph Kalo

University of Amsterdam

Kubilay Atasu

Delft University of Technology

Christopher Morris

RWTH Aachen University

Schedule

14:00 – 14:05

Opening remarks

Welcome and introduction by the organizers
14:05 – 14:45

Jan-Christoph Kalo

Explicit Knowledge Models
14:45 – 15:25

Kubilay Atasu

Financial Crime Detection: Decentralization, Privacy, Robustness, and Utility
15:25 – 15:45
Lightning talks
1. Addressing Structural Heterogeneity with Ontology-Aware Message Passing
  
  Shervin Mehryar
2. From Flow to Features: Directed Learnable Positional Encodings for Anti-Money Laundering
  
  Fatemeh Gholamzadeh Nasrabadi, Alireza Khayatian, Michael Werner, and Marcel Worring
3. MEGA: Message Passing Neural Networks for Multigraphs with EdGe Attributes
  
  Halil Cagri Bilgi and Kubilay Atasu
4. Translating Latent Representations for Money Laundering Detection
  
  Ramon Rico, Ioana Hulpus, Stan Leisink, Boyang Zhao, and Yannis Velegrakis
15:45 – 16:45

Poster session
16:45 – 17:25

Christopher Morris

Which Algorithms can Graph Neural Networks Learn?
17:25 – 17:30

Closing remarks

Accepted abstracts

The following extended abstracts will be presented during the poster session.

01

Addressing Structural Heterogeneity with Ontology-Aware Message Passing

Shervin Mehryar
02

From Flow to Features: Directed Learnable Positional Encodings for Anti-Money Laundering

Fatemeh Gholamzadeh Nasrabadi, Alireza Khayatian, Michael Werner, and Marcel Worring
03

MEGA: Message Passing Neural Networks for Multigraphs with EdGe Attributes

Halil Cagri Bilgi and Kubilay Atasu
04

Translating Latent Representations for Money Laundering Detection

Ramon Rico, Ioana Hulpus, Stan Leisink, Boyang Zhao, and Yannis Velegrakis

Call for abstracts

We invite extended abstracts on learning and reasoning with graphs, spanning foundational methods and real-world applications, to be presented during the poster session.

Authors are invited to submit an extended abstract of up to 2 pages (excluding references) using the Springer Lecture Notes in Computer Science (LNCS) template. The template is available on Overleaf.

Submissions will be lightly reviewed for relevance to the workshop, and they may describe previously published work, work in progress, or ongoing research.