AGENT99 TRAINER: DESIGNING A WEB-BASEDMULTIMEDIA TRAINING SYSTEM FOR DECEPTIONDETECTION KNOWLEDGE TRANSFER

摘要

Although detection of malicious deception is important to prevention of possible harm to individuals ororganizations, human deception detection accuracy is poor (around 40%-60%) because people lack deceptiondetection knowledge. Therefore it is important to create strategies and tools for making effective transfer ofdeception detection knowledge readily available to the masses. This paper describes our research efforts todesign and implement a Web-based multimedia training system, Agent99 Trainer, to provide effective deceptiondetection knowledge transfer. Agent99 Trainer captures experts’ knowledge on multimedia, presents it tolearners through an integrated interface, provides practice opportunities by allowing viewing of real lifeexamples, and gives immediate and elaborate feedback. Furthermore all the components of Agent99 Trainerare integrated together to achieve the best possible performance. Our initial experiment shows that trainingusing Agent99 Trainer improves human deception detection accuracy and that the training is as effective aslecture-based training. In addition, our usability study shows that Agent99 Trainer is easy to use and supportsthe system design from a subjective perspective.