v231: Proceedings of Learning on Graphs 2023

2026-06-10 · Source: Proceedings of Machine Learning Research · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Mathematics & Computational Sciences · Depth: Expert, medium

Summary

The Second Learning on Graphs Conference, held virtually from November 27-30, 2023, showcased a wide array of research in graph-based machine learning. Key advancements included novel Graph Neural Networks (GNNs) addressing challenges such as over-smoothing, over-squashing, and performance discrepancies related to graph homophily. Papers explored diverse graph representation learning techniques, including cycle invariant positional encoding, spectral subgraph localization, and effective multidimensional persistence. Significant contributions also focused on neural algorithmic reasoning, generative models for protein and molecule structures, and applications in fluid dynamics and combinatorial optimization. The proceedings highlighted new software packages like "PyTorch Geometric Signed Directed" and benchmarks such as "SURF" for GNNs in fluid dynamics.

Key takeaway

The Learning on Graphs Conference (LoG 2023) proceedings showcase cutting-edge research advancing Graph Neural Networks (GNNs) and graph representation learning. Key contributions address challenges like over-smoothing, homophily, and algorithmic reasoning, introducing novel techniques for multi-relational graphs, cycle-invariant positional encoding, and hypergraph learning. This collection offers critical insights and open-source tools for AI/ML researchers and practitioners developing robust graph-based solutions across domains like fluid dynamics, molecule generation, and combinatorial optimization.

Topics

• Graph Neural Networks
• Graph Representation Learning
• Neural Algorithmic Reasoning
• Graph Generative Models
• Hypergraph Neural Networks

Code references

• josefhoppe/edge-flow-cell-complexes
• DJayalath/gnn-call-stack
• lukasjf/gwac
• GraphEoM/GSCAN
• mirjanic/nar-latent-spaces

Best for: AI Scientist, Machine Learning Engineer, Research Scientist

Related on AIssential

• v269: Proceedings of Learning on Graphs 2025 — The conference highlights diverse advancements in graph learning, spanning theoretical foundations, architectural innovations, and real-world applications.
• v198: Proceedings of Learning on Graphs 2022 — The First Learning on Graphs Conference, held virtually from December 9-12, 2022, compiled a significant body of research in graph machine learning, edited by Bastian Rieck and Razvan Pascanu.
• Graph Neural Networks for Predicting Solvability of Finite Groups — GNNs can classify finite group solvability using structural graph representations like Cayley graphs.
• A Topological Characterization of Graph Neural Networks via Stochastic Block Model Embeddings on the n-Sphere — A topological framework characterizes GNNs via SBM embeddings on an n-sphere, creating low-dimensional "fingerprints" for model comparison and transfer learning.
• A Gentle Introduction to Graph Neural Networks — GNNs leverage graph structures to model relationships, enabling predictions across diverse data types and tasks.
• Multi-hop Relational Contrastive Learning: Extending Spatial Contrastive Pre-training Beyond Pairwise Relations — Multi-hop relational contrastive learning captures complex spatial dependencies beyond pairwise relations, improving scene understanding.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Proceedings of Machine Learning Research.