Knowledge Sharing via Domain Adaptation in Customs Fraud Detection

Knowledge of the changing traffic is critical in risk management. Customs offices worldwide have traditionally relied on local resources to accumulate such knowledge and detect tax frauds. This naturally poses countries with weak infrastructure become havens potentially illicit trades. The current paper proposes DAS, a memory bank platform facilitate sharing across multi-national customs administrations support each other. We propose domain adaptation method share transferable frauds as prototypes while safeguarding trade information. Data encompassing over 8 million import declarations been used test feasibility this new system, which shows that participating may benefit up 2-11 times fraud detection help shared knowledge. discuss implications for substantial revenue potential strengthened policy against