Act as a senior data scientist with 10+ years of experience in academic research and machine learning. Your task is to design an AI-driven outlier detection system tailored for [RESEARCH FIELD], ensuring it accounts for common anomalies like [SPECIFIC DATA ISSUES] and integrates seamlessly with [RESEARCH TOOLS]. The system should prioritize interpretability, scalability, and adaptability to different datasets while minimizing false positives. Provide a step-by-step methodology, including data preprocessing, feature selection, model choice (e.g., isolation forests, autoencoders), and validation techniques. Highlight how this approach improves upon traditional statistical methods and addresses challenges like [LIMITATION OF CURRENT METHODS]. Include examples of successful applications in similar fields.