---
title: Transaction Fraud Detection via Spatial-Temporal-Aware Graph Transformer
url: https://www.ml-quant.com/papers/arxiv/2307.05121/
site: ML-Quant (https://www.ml-quant.com)
updated: 2026-09-26
license: Summaries CC BY 4.0; links go to the original sources
index: https://www.ml-quant.com/llms.txt
identifier: arXiv:2307.05121
source_url: https://arxiv.org/abs/2307.05121
featured: 2023-07-12
citations: 10
topic: ML & AI Methods
---


# Transaction Fraud Detection via Spatial-Temporal-Aware Graph Transformer

A new graph neural network, STA-GT, is proposed for transaction fraud detection, effectively learning and incorporating spatial-temporal and global information.

- Source: https://arxiv.org/abs/2307.05121
- Identifier: arXiv:2307.05121
- Released: 2023-07-11
- First featured: Quant Letter No. 7 (2023-07-12): https://www.ml-quant.com/issues/2023-07-12/
- Citations (Semantic Scholar): 10
- Published in: not yet
- Topic: ML & AI Methods

## Related

- [Mamba: Linear-Time Sequence Modeling with Selective State Spaces](https://www.ml-quant.com/papers/arxiv/2312.00752/): Sequence Modeling: Mamba, a neural network architecture that doesn't use attention or MLP blocks, provides faster inference and better performance in language, audio, and genomics than Transformers.
- [Graph Mamba: Towards Learning on Graphs with State Space Models](https://www.ml-quant.com/papers/arxiv/2402.08678/): Graph Mamba Networks, a new type of Graph Neural Networks, have been introduced, which achieve excellent performance in various benchmark datasets despite lower computational cost.
- [Edge Directionality Improves Learning on Heterophilic Graphs](https://www.ml-quant.com/papers/arxiv/2305.10498/): The study presents Directed Graph Neural Network (Dir-GNN), a new deep learning framework for directed graphs that surpasses traditional models in heterophilic benchmarks.
- [Anomaly Detection in Global Financial Markets with Graph Neural Networks and Nonextensive Entropy](https://www.ml-quant.com/papers/arxiv/2308.02914/): A study using Graph Neural Networks to identify anomalies in global financial markets found that the interconnected structure of highly correlated assets decreases during a crisis, with the number of anomalies varying based on the crisis stage.
- [GNN-VPA: A Variance-Preserving Aggregation Strategy for Graph Neural Networks](https://www.ml-quant.com/papers/arxiv/2403.04747/): A new variance-preserving aggregation function for Graph Neural Networks is proposed, which could lead to more efficient and self-normalizing networks.
- [BloomGML: Graph Machine Learning through the Lens of Bilevel Optimization](https://www.ml-quant.com/papers/arxiv/2403.04763/): Graph Machine Learning via Bilevel Optimization: The paper views graph learning techniques as special cases of bilevel optimization, introducing a new class of energy functions for graph neural network layers, and showcasing the versatility of this approach through empirical results.