English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

The application of graph representation learning techniques to the area offinancial risk management (FRM) has attracted significant attention recently.However, directly modeling transaction networks using graph neural modelsremains challenging: Firstly, transaction networks are directed multigraphs bynature, which could not be properly handled with most of the currentoff-the-shelf graph neural networks (GNN). Secondly, a crucial problem in FRMscenarios like anti-money laundering (AML) is to identify risky transactionsand is most naturally cast into an edge classification problem with richedge-level features, which are not fully exploited by the prevailing GNN designthat follows node-centric message passing protocols. In this paper, we presenta systematic investigation of design aspects of neural models over directedmultigraphs and develop a novel GNN protocol that overcomes the abovechallenges via efficiently incorporating directional information, as well asproposing an enhancement that targets edge-related tasks using a novel messagepassing scheme over an extension of edge-to-node dual graph. A concrete GNNarchitecture called GRANDE is derived using the proposed protocol, with severalfurther improvements and generalizations to temporal dynamic graphs. We applythe GRANDE model to both a real-world anti-money laundering task and publicdatasets. Experimental evaluations show the superiority of the proposed GRANDEarchitecture over recent state-of-the-art models on dynamic graph modeling anddirected graph modeling.

GRANDE: a neural model over directed multigraphs with application to anti-money laundering