Multimodal Learning for Fraud Detection in Last-Mile Delivery

Created using ChatSlide

The presentation explores the advent of multimodal learning for enhancing fraud detection in last-mile delivery services, addressing the drawbacks of conventional GPS methods. Focused on the Multimodal Fraud Detection (MFDD) framework, it outlines key features such as multimodal embedding extraction and cross-modal fusion techniques. Core contributions highlight its pioneering role in multimodal fraud studies and effective platform deployment. Utilising real-world data from JD Logistics,...