machine-learning-pipeline

The paper evaluates a configurable multi-agent system (nMAS) for extracting structured oncology data from fragmented clinical documents, achieving high performance compared to a baseline model.