OncoAgent: A Dual-Layer Multi-Agent Framework for Privacy-Preserving Tumor Clinical Decision Support

robot
Abstract generation in progress

AIMPACT News, May 10 (UTC+8), the research team released the open-source tumor clinical decision support system OncoAgent. The system uses a dual-layer multi-agent framework, combining LangGraph topology with a four-stage Corrective RAG process to retrieve more than 70 authoritative clinical guidelines. Based on query complexity, the system routes tasks to either a speed-optimized model with 9B parameters or a deep reasoning model with 27B parameters; both are fine-tuned with QLoRA on AMD MI300X hardware using a dataset of more than 260k cases. The system enforces a strict policy of zero protected health information and ensures security through a three-layer reflective safety validator, supporting fully local deployment to protect patient data sovereignty. (Source: AiHot)

View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • Comment
  • Repost
  • Share
Comment
Add a comment
Add a comment
No comments
  • Pin