Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Video

Updated: Apr 8, 2026

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

1.3K

FedKGC: Federated Knowledge-Grounded Calibration Framework for Hallucination Mitigation in Medical LLMs.

Chinmay Chakraborty, Soufiane Ben Othman, Manisha Guduri

    IEEE Journal of Biomedical and Health Informatics
    |April 6, 2026
    PubMed
    Summary

    Federated Knowledge-Grounded Calibration (FedKGC) reduces AI hallucinations in clinical settings by using federated learning and knowledge grounding. This privacy-preserving approach enables collaborative training of trustworthy medical AI across institutions.

    Related Concept Videos

    You might also read

    Related Articles

    Articles linked to this work by shared authors, journal, and citation graph.

    Sort by
    Same author

    Quantum-enhanced privacy aggregation for healthcare monitoring in wireless body area networks.

    Scientific reports·2026
    Same author

    Uncovering various neuronal responses in a fractional-order generalized HR system.

    Neural networks : the official journal of the International Neural Network Society·2026
    Same author

    Scalable privacy-preserving data analytics for IoMT via FHE and zk-SNARK-enabled edge aggregation.

    Scientific reports·2026
    Same author

    Cell-Level Free Cervical Lesion Detection in Cytology Images Via Weakly Supervised Self-Correction.

    IEEE journal of biomedical and health informatics·2025
    Same author

    Quantum-entangled neuro-symbolic swarm federation for privacy-preserving IoMT-driven multimodal healthcare.

    Scientific reports·2025
    Same author

    SQAHO-IoMT: A Post-Quantum Secure and AI-Resilient Aggregation Framework for Smart Healthcare Systems.

    IEEE journal of biomedical and health informatics·2025

    Area of Science:

    • Artificial Intelligence
    • Medical Informatics
    • Machine Learning

    Background:

    • Large Language Models (LLMs) show promise for clinical decision support and documentation.
    • Widespread adoption is hindered by AI hallucination risks and data privacy regulations.

    Purpose of the Study:

    • To introduce FedKGC, a novel framework combining Federated Learning (FL) with retrieval-augmented calibration.
    • To mitigate hallucination and ensure privacy in distributed medical AI training.

    Main Methods:

    • FedKGC employs a dual-phase strategy: Knowledge-Grounded Loss (L_KG) for local fine-tuning and differentially private metadata for global aggregation.
    • A Global Conflict Map is constructed to down-weight contradictory concepts during aggregation.

    More Related Videos

    Evidence-based Knowledge Synthesis and Hypothesis Validation: Navigating Biomedical Knowledge Bases via Explainable AI and Agentic Systems
    05:47

    Evidence-based Knowledge Synthesis and Hypothesis Validation: Navigating Biomedical Knowledge Bases via Explainable AI and Agentic Systems

    Published on: June 13, 2025

    1.9K

    Related Experiment Videos

    Last Updated: Apr 8, 2026

    Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
    03:14

    Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

    Published on: December 6, 2024

    1.3K
    Evidence-based Knowledge Synthesis and Hypothesis Validation: Navigating Biomedical Knowledge Bases via Explainable AI and Agentic Systems
    05:47

    Evidence-based Knowledge Synthesis and Hypothesis Validation: Navigating Biomedical Knowledge Bases via Explainable AI and Agentic Systems

    Published on: June 13, 2025

    1.9K

    Main Results:

    • FedKGC achieved a 28% reduction in hallucination rates compared to standard federated baselines.
    • The framework adheres to (ε, δ)-Differential Privacy guarantees on MIMIC-IV and MedHallu datasets.

    Conclusions:

    • High-fidelity, hallucination-resistant medical AI can be collaboratively trained across institutions without compromising patient privacy.
    • FedKGC facilitates trustworthy distributed medical intelligence.