Banks often uncover new fraud cases by tracing connections from previously confirmed ones. In practice, this process faces two intertwined challenges: audits delay label confirmation, and fraud groups continually evolve their behaviors. These challenges manifest as label delay (LD) and concept drift (CD), which jointly degrade fraud detection. Errors introduced by delayed labels can propagate along network links, further amplifying the impact of concept drift. We propose a Data-Centric Dual-Layer Adversarial Framework (DDAF) that jointly addresses LD and CD by improving data robustness in a backbone-agnostic way. DDAF introduces a counterfactual attribution module to disentangle and quantify the individual and interaction effects of LD and CD. Guided by such attributions, we construct adversarial training graphs at two levels: node-level feature perturbations to mitigate delay-induced bias, and community-level structural perturbations to adapt to evolving group behavior. This dual-layer design coordinates correction across node and community scales, reducing error accumulation over time. On public datasets, DDAF improves AUC by 3.4 points and F1 by 3.5 points over strong baselines. In a large commercial bank deployment, DDAF identified 1,449 previously unknown fraud accounts, and expanded risk coverage in daily operations.

Real-world networks containing hyper-relational facts can be represented by Hyper-relational Knowledge Graphs (HKGs), which contain rich semantics compared to hypergraphs. In an HKG, multiple entities with a shared property playing different roles form a hyperedge. To mitigate information loss in HKG representation learning, it is important to integrate entity-level roles and hyperedge-level properties. However, such multi-grained semantics are often inadequately learned, as conventional message passing only operates along one granularity, i.e, hyperedges. We propagate information between entities and hyperedges to aggregate entity-level roles at the hyperedge level for message passing. In this paper, we propose an HKG representation learning method based on Multi-granularity Semantic aware Message Passing (HyperSMP), preserving both semantic and structural information. Specifically, we design a fine-grained aggregation layer in HyperSMP to aggregate different roles of entities in hyperedge embeddings using an entity-level attention mechanism. Based on hyperedge embeddings, we propose a granularity-aligned propagation layer that recursively propagates information from hyperedges to entities and properties, capturing message passing paths in and out of hyperedges, respectively. As a result, multi-grained semantics are learned through unification at the hyperedge level to adapt to the message passing mechanism, and the high-order structure is captured through entity information exchange channeled via hyperedge embeddings. Further, HyperSMP discovers multi-hop associations among entities by stacking the above layers. Experimental results on public datasets show that HyperSMP outperforms state-of-the-art methods by at least 9.66% in F1 score.

Disclosing Actual Controllers (ACs) of a company has been the basis for financial risk governance. A shareholder in a winning stable coalition, where members make consistent decisions and win in votes, is considered an AC. However, existing methods fail to discover stable coalitions due to the ignorance of various relations other than the shareholding relation among shareholders, such as kinship, subsidiary and so on. Moreover, the above relations form a large-scale equity network, which brings challenges for efficiently identifying winning stable coalitions.