Details

Autor(en) / Beteiligte

• Li, Zilinghan
• He, Shilan
• Chaturvedi, Pranshu
• Kindratenko, Volodymyr
• Huerta, Eliu A
• Kim, Kibaek
• Madduri, Ravi

Titel

Secure Federated Learning Across Heterogeneous Cloud and High-Performance Computing Resources - A Case Study on Federated Fine-tuning of LLaMA 2

Ist Teil von

• Computing in science & engineering, 2024, p.1-7

Ort / Verlag

IEEE

Erscheinungsjahr

2024

Quelle

IEEE Xplore

Beschreibungen/Notizen

• Federated learning enables multiple data owners to collaboratively train robust machine learning models without transferring large or sensitive local datasets by only sharing the parameters of the locally trained models. In this paper, we elaborate on the design of our Advanced Privacy-Preserving Federated Learning (APPFL) framework, which streamlines end-to-end secure and reliable federated learning experiments across cloud computing facilities and high-performance computing resources by leveraging Globus Compute, a distributed function as a service platform, and Amazon Web Services. We further demonstrate the use case of APPFL in finetuning a LLaMA 2 7B model using several cloud resources and supercomputers.