DuaLK integrates medical knowledge with EHR data to create AI diagnostic systems that mirror physicians' stepwise reasoning process.

• Combines lab test signals with comprehensive medical knowledge graphs
• Implements a two-stage approach: knowledge-guided pretraining followed by fine-tuning
• Achieves superior performance in diagnostic reasoning compared to traditional models
• Provides interpretable reasoning paths that align with clinical decision-making

This research bridges the gap between data-driven AI and human clinical expertise, potentially reducing diagnostic errors and improving patient outcomes in healthcare settings.

Bridging Stepwise Lab-Informed Pretraining and Knowledge-Guided Learning for Diagnostic Reasoning