PALISADE-X supports real-world, high-impact applications of privacy-preserving federated learning across multiple domains. The use cases below highlight how APPFL, APPFLx, AIDRIN, and CADRE enable collaborative AI without centralizing sensitive data.
Biomedical data are often highly sensitive, institutionally siloed, and governed by strict privacy and compliance requirements. PALISADE-X enables collaborative AI development across healthcare and research institutions while ensuring that patient data remain local and protected.
Recent work has demonstrated the feasibility and value of federated learning for large-scale biomedical AI, particularly in medical imaging and vision-based diagnostics. By training models across multiple sites without pooling data, federated approaches improve model generalizability while preserving privacy and institutional control.
Key biomedical use cases include:
Federated training of medical imaging models across hospitals and research centers
Collaborative development of AI models for disease detection and progression analysis
Cross-institution learning on heterogeneous datasets without data sharing
Reproducible and policy-aware AI pipelines for regulated biomedical environments
These approaches reduce bias introduced by single-site datasets and enable broader participation in AI-driven biomedical discovery.
PALISADE-X supports hands-on biomedical federated learning through the AI-READI ARVO tutorial, which demonstrates end-to-end federated model training using real-world biomedical workflows. The tutorial focuses on practical execution, experiment tracking, and privacy-preserving collaboration across institutions.
Related resources
Biomedical federated learning study:
https://www.sciencedirect.com/science/article/pii/S2001037024004239
AI-READI ARVO Federated Learning Tutorial (APPFL):
https://github.com/APPFL/APPFL/tree/main/examples/notebook_tutorials/aireadi_arvo_course
Modern power grids generate large volumes of operational and sensor data that are distributed across utilities, regions, and infrastructure operators. Centralizing these data is often infeasible due to security, policy, and operational constraints.
PALISADE-X enables federated AI for grid and energy systems, allowing stakeholders to collaboratively train models while maintaining local control over sensitive infrastructure data.
Federated learning enables:
Collaborative training of grid foundation models across utilities and regions
Privacy-preserving analysis of operational grid data
Improved resilience and generalization of AI models for grid monitoring and forecasting
Scalable execution across cloud and high-performance computing environments
These capabilities support AI-driven grid reliability, optimization, and resilience without exposing raw operational data.
PALISADE-X provides practical examples demonstrating how APPFL can be applied to grid datasets using federated learning workflows. These tutorials illustrate data partitioning, federated coordination, and scalable training across distributed environments.
Related resources
Grid federated learning tutorial (APPFL):
https://github.com/APPFL/APPFL/tree/main/examples/notebook_tutorials/grid