Knowledge Guided

Knowledge-guided machine learning (KGML) integrates domain-specific knowledge with machine learning models to improve prediction accuracy, explainability, and adherence to scientific principles. Current research focuses on developing methods to effectively incorporate diverse knowledge sources, such as knowledge graphs and expert-defined rules, into various ML architectures, including semi-supervised learning and reinforcement learning frameworks. This approach is proving valuable across diverse fields, from environmental modeling (e.g., wildfire prediction) and video quality assessment to drug repurposing and natural language processing, enhancing the reliability and interpretability of AI-driven insights.