Knowledge-Integrated Representation Learning for Crypto Anomaly Detection under Extreme Label Scarcity; Relational Domain-Logic Integration with Retrieval-Grounded Context and Path-Level Explanations

Paper: arXiv 2601.12839

Authors: Gyuyeon Na, Minjung Park, Soyoun Kim, Jungbin Shin, Sangmi Chai

Abstract

Detecting anomalous trajectories in decentralized crypto networks is fundamentally challenged by extreme label scarcity and the adaptive evasion strategies of illicit actors. While Graph Neural Networks (GNNs) effectively capture local structural patterns, they struggle to internalize multi hop, logic driven motifs such as fund dispersal and layering that characterize sophisticated money laundering, limiting their forensic accountability under regulations like the FATF Travel Rule. To address this limitation, we propose Relational Domain Logic Integration (RDLI), a framework that embeds expert derived heuristics as differentiable, logic aware latent signals within representation learning. Unlike static rule based approaches, RDLI enables the detection of complex transactional flows that evade standard message passing. To further account for market volatility, we incorporate a Retrieval Grounded Context (RGC) module that conditions anomaly scoring on regulatory and macroeconomic context, mitigating false positives caused by benign regime shifts. Under extreme label scarcity (0.01%), RDLI outperforms state of the art GNN baselines by 28.9% in F1 score. A micro expert user study further confirms that RDLI path level explanations significantly improve trustworthiness, perceived usefulness, and clarity compared to existing methods, highlighting the importance of integrating domain logic with contextual grounding for both accuracy and explainability.

Complexity vs Empirical Score

• Math Complexity: 4.5/10
• Empirical Rigor: 7.5/10
• Quadrant: Street Traders — practical and empirical, lighter on theory

Why this score: The paper presents a pragmatic blend of graph neural networks, LLM-generated expert knowledge graphs, and retrieval-augmented context to address extreme label scarcity in crypto anomaly detection. Its empirical validation includes both quantitative benchmarks under severe scarcity (0.01%) and a human-in-the-loop expert user study.

Research Flowchart

  flowchart TD
    A[Goal: Detect Crypto Anomalies under Extreme Label Scarcity 0.01%] --> B[Inputs: On-Chain Transactions, Regulatory/Macro Context, Expert Logic]
    B --> C[Relational Domain Logic Integration: Embed Differentiable Heuristics into GNN]
    B --> D[Retrieval-Grounded Context Module: Condition Scoring on Market Regimes]
    C & D --> E[Representation Learning & Anomaly Scoring]
    E --> F[Outcomes: +28.9% F1 Score Over GNN Baselines]
    E --> G[Outcomes: Path-Level Explanations for Forensic Accountability]

Cited References (47)

1. Credit Card Transactions Dataset (2023)
2. CSGNN: Contrastive Self-Supervised Graph Neural Network for Molecular Interaction Prediction. (2021)
3. Robust Embedding with Multi-Level Structures for Link Prediction. (2019)
4. Structure and Interpretation of Computer Programs (1985)
5. Inductive representation learning on large graphs (2017)
6. Cooperative Joint Attentive Network for Patient Outcome Prediction on Irregular Multi-Rate Multivariate Health Data. (2021)
7. Reading selectively via Binary Input Gated Recurrent Unit. (2019)
8. Visual web information extraction with lixto (2001)

0 of 47 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2518 linked papers.