Back
13677_leap_local_ect_based_lea
Conference paper (in proceedings)

LEAP: Local ECT-Based Learnable Positional Encodings for Graphs

BP2-STS

  • 2026
Published in:
  • International Conference on Learning Representations (ICLR). - 2026
Topological data analysis
Topological deep learning
Geometric deep learning
Euler characteristic transform
Euler characteristic
English Graph neural networks (GNNs) largely rely on the message-passing paradigm, where nodes iteratively aggregate information from their neighbors. Yet, standard message passing neural networks (MPNNs) face well-documented theoretical and practical limitations. Graph positional encoding (PE) has emerged as a promising direction to address these limitations. The Euler Characteristic Transform (ECT) is an efficiently computable geometric–topological invariant that characterizes shapes and graphs. In this work, we combine the differentiable approximation of the ECT (DECT) and its local variant (l-ECT) to propose LEAP, a new end-to-end trainable local structural PE for graphs. We evaluate our approach on multiple real-world datasets as well as on a synthetic task designed to test its ability to extract topological features. Our results underline the potential of LEAP-based encodings as a powerful component for graph representation learning pipelines.
Faculty
Faculté des sciences et de médecine
Department
Département d'Informatique
Language
  • English
Classification
Computer science and technology
Other electronic version

Published version

License
License undefined
Open access status
green
Persistent URL
https://folia.unifr.ch/unifr/documents/336222
Statistics
Document views: 3 File downloads:
  • 13677_leap_local_ect_based_lea: 0