Subgraphormer: Subgraph GNNs meet Graph Transformers

G Bar-Shalom, B Bevilacqua… - NeurIPS 2023 Workshop …, 2023 - openreview.net
In the realm of Graph Neural Network (GNNs), two intriguing research directions have
recently emerged: Subgraph GNNs and Graph Transformers. These approaches have …

BEVSeg2GTA: Joint Vehicle Segmentation and Graph Neural Networks for Ego Vehicle Trajectory Prediction in Bird's-Eye-View

S Sharma, A Das, G Sistu, M Halton, C Eising - IEEE Access, 2024 - ieeexplore.ieee.org
Predicting the trajectory of the ego vehicle is a critical task for autonomous vehicles. Even
though traffic regulations have defined boundaries, various behaviors of the agents in real …

Towards Graph Foundation Models: A Study on the Generalization of Positional and Structural Encodings

BJ Franks, M Eliasof, S Cantürk, G Wolf… - arXiv preprint arXiv …, 2024 - arxiv.org
Recent advances in integrating positional and structural encodings (PSEs) into graph neural
networks (GNNs) have significantly enhanced their performance across various graph …

Disentangled Structural and Featural Representation for Task-Agnostic Graph Valuation

A Falahati, MM Amiri - arXiv preprint arXiv:2408.12659, 2024 - arxiv.org
With the emergence of data marketplaces, the demand for methods to assess the value of
data has increased significantly. While numerous techniques have been proposed for this …

Learning Graph Quantized Tokenizers for Transformers

L Wang, K Hassani, S Zhang, D Fu, B Yuan… - arXiv preprint arXiv …, 2024 - arxiv.org
Transformers serve as the backbone architectures of Foundational Models, where a domain-
specific tokenizer helps them adapt to various domains. Graph Transformers (GTs) have …

ExGRG: Explicitly-Generated Relation Graph for Self-Supervised Representation Learning

M Naseri, M Biparva - arXiv preprint arXiv:2402.06737, 2024 - arxiv.org
Self-supervised Learning (SSL) has emerged as a powerful technique in pre-training deep
learning models without relying on expensive annotated labels, instead leveraging …

On the Utilization of Unique Node Identifiers in Graph Neural Networks

M Bechler-Speicher, M Eliasof, CB Schönlieb… - arXiv preprint arXiv …, 2024 - arxiv.org
Graph neural networks have inherent representational limitations due to their message-
passing structure. Recent work has suggested that these limitations can be overcome by …

Range-aware Positional Encoding via High-order Pretraining: Theory and Practice

VA Nguyen, NK Ngo, TS Hy - arXiv preprint arXiv:2409.19117, 2024 - arxiv.org
Unsupervised pre-training on vast amounts of graph data is critical in real-world applications
wherein labeled data is limited, such as molecule properties prediction or materials science …

Optimizing Ego Vehicle Trajectory Prediction: The Graph Enhancement Approach

S Sharma, A Singh, G Sistu, M Halton… - arXiv preprint arXiv …, 2023 - arxiv.org
Predicting the trajectory of an ego vehicle is a critical component of autonomous driving
systems. Current state-of-the-art methods typically rely on Deep Neural Networks (DNNs) …

On the expressivity and sample complexity of node-individualized graph neural networks

P Pellizzoni, TH Schulz, D Chen… - The Thirty-eighth Annual …, 2024 - openreview.net
Graph neural networks (GNNs) employing message passing for graph classification are
inherently limited by the expressive power of the Weisfeiler-Leman (WL) test for graph …