LEAP: Local ECT-Based Learnable Positional Encodings for Graphs
This work addresses graph representation learning challenges for researchers and practitioners, but it appears incremental as it builds on existing positional encoding and ECT methods.
The authors tackled the limitations of standard message-passing neural networks (MPNNs) in graph neural networks (GNNs) by proposing LEAP, a trainable local structural positional encoding based on the Euler Characteristic Transform (ECT), and demonstrated its effectiveness on multiple real-world datasets and a synthetic task.
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 ($\ell$-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.