Kernelized Edge Attention: Addressing Semantic Attention Blurring in Temporal Graph Neural Networks

• URL: http://arxiv.org/abs/2602.00596v1
• Date: Sat, 31 Jan 2026 08:22:35 GMT
• Title: Kernelized Edge Attention: Addressing Semantic Attention Blurring in Temporal Graph Neural Networks
• Authors: Govind Waghmare, Srini Rohan Gujulla Leel, Nikhil Tumbde, Sumedh B G, Sonia Gupta, Srikanta Bedathur,
• Abstract summary: This paper introduces KEAT, a novel attention formulation that modulates edge features using a family of continuous-time kernels.<n>It achieves up to 18% MRR improvement over the recent DyGFormer and 7% over TGN on link prediction tasks.
• Score: 7.383288419236205
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Temporal Graph Neural Networks (TGNNs) aim to capture the evolving structure and timing of interactions in dynamic graphs. Although many models incorporate time through encodings or architectural design, they often compute attention over entangled node and edge representations, failing to reflect their distinct temporal behaviors. Node embeddings evolve slowly as they aggregate long-term structural context, while edge features reflect transient, timestamped interactions (e.g. messages, trades, or transactions). This mismatch results in semantic attention blurring, where attention weights cannot distinguish between slowly drifting node states and rapidly changing, information-rich edge interactions. As a result, models struggle to capture fine-grained temporal dependencies and provide limited transparency into how temporal relevance is computed. This paper introduces KEAT (Kernelized Edge Attention for Temporal Graphs), a novel attention formulation that modulates edge features using a family of continuous-time kernels, including Laplacian, RBF, and learnable MLP variant. KEAT preserves the distinct roles of nodes and edges, and integrates seamlessly with both Transformer-style (e.g., DyGFormer) and message-passing (e.g., TGN) architectures. It achieves up to 18% MRR improvement over the recent DyGFormer and 7% over TGN on link prediction tasks, enabling more accurate, interpretable and temporally aware message passing in TGNNs.

Related papers

• Full-History Graphs with Edge-Type Decoupled Networks for Temporal Reasoning [16.53173953073833]
  We introduce a full-history graph that instantiates one node for every entity at every time step.<n>We evaluate it on driverintention prediction (Waymo) and Bitcoin fraud detection (Elliptic++)<n>These gains demonstrate the benefit representing structural and temporal relations as distinct edges in a single graph.
  arXiv  Detail & Related papers  (2025-08-05T09:29:07Z)
• Adaptive Sparsified Graph Learning Framework for Vessel Behavior Anomalies [3.3711670942444014]
  Graph neural networks have emerged as a powerful tool for learning precise interactions.<n>Our method introduces an innovative graph representation where edges are modeled as timestamp nodes.<n>This setup is extended to construct a multi-ship graph that captures spatial interactions while preserving graph sparsity.
  arXiv  Detail & Related papers  (2025-02-20T02:01:40Z)
• Beyond Message Passing: Neural Graph Pattern Machine [50.78679002846741]
  We introduce the Neural Graph Pattern Machine (GPM), a novel framework that bypasses message passing by learning directly from graph substructures.<n>GPM efficiently extracts, encodes, and prioritizes task-relevant graph patterns, offering greater expressivity and improved ability to capture long-range dependencies.
  arXiv  Detail & Related papers  (2025-01-30T20:37:47Z)
• Improving Graph Neural Networks by Learning Continuous Edge Directions [0.0]
  Graph Neural Networks (GNNs) traditionally employ a message-passing mechanism that resembles diffusion over undirected graphs.<n>Our key insight is to assign fuzzy edge directions to the edges of a graph so that features can preferentially flow in one direction between nodes.<n>We propose a general framework, called Continuous Edge Direction (CoED) GNN, for learning on graphs with fuzzy edges.
  arXiv  Detail & Related papers  (2024-10-18T01:34:35Z)
• Temporal Aggregation and Propagation Graph Neural Networks for Dynamic Representation [67.26422477327179]
  Temporal graphs exhibit dynamic interactions between nodes over continuous time. We propose a novel method of temporal graph convolution with the whole neighborhood. Our proposed TAP-GNN outperforms existing temporal graph methods by a large margin in terms of both predictive performance and online inference latency.
  arXiv  Detail & Related papers  (2023-04-15T08:17:18Z)
• TPGNN: Learning High-order Information in Dynamic Graphs via Temporal Propagation [7.616789069832552]
  We propose a temporal propagation-based graph neural network, namely TPGNN. Propagator propagates messages from anchor node to temporal neighbors within $k$-hop, and then simultaneously update the state of neighborhoods. To prevent over-smoothing, the model compels the messages from $n$-hop neighbors to update the $n$-hop memory vector preserved on the anchor.
  arXiv  Detail & Related papers  (2022-10-03T18:39:07Z)
• Dynamic Graph Message Passing Networks for Visual Recognition [112.49513303433606]
  Modelling long-range dependencies is critical for scene understanding tasks in computer vision. A fully-connected graph is beneficial for such modelling, but its computational overhead is prohibitive. We propose a dynamic graph message passing network, that significantly reduces the computational complexity.
  arXiv  Detail & Related papers  (2022-09-20T14:41:37Z)
• Spatio-Temporal Joint Graph Convolutional Networks for Traffic Forecasting [75.10017445699532]
  Recent have shifted their focus towards formulating traffic forecasting as atemporal graph modeling problem. We propose a novel approach for accurate traffic forecasting on road networks over multiple future time steps.
  arXiv  Detail & Related papers  (2021-11-25T08:45:14Z)
• Gated Graph Recurrent Neural Networks [176.3960927323358]
  We introduce Graph Recurrent Neural Networks (GRNNs) as a general learning framework for graph processes. To address the problem of vanishing gradients, we put forward GRNNs with three different gating mechanisms: time, node and edge gates. The numerical results also show that GRNNs outperform GNNs and RNNs, highlighting the importance of taking both the temporal and graph structures of a graph process into account.
  arXiv  Detail & Related papers  (2020-02-03T22:35:14Z)
• EdgeNets:Edge Varying Graph Neural Networks [179.99395949679547]
  This paper puts forth a general framework that unifies state-of-the-art graph neural networks (GNNs) through the concept of EdgeNet. An EdgeNet is a GNN architecture that allows different nodes to use different parameters to weigh the information of different neighbors. This is a general linear and local operation that a node can perform and encompasses under one formulation all existing graph convolutional neural networks (GCNNs) as well as graph attention networks (GATs)
  arXiv  Detail & Related papers  (2020-01-21T15:51:17Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.