Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces a workflow framework designed for asynchronous human-AI collaboration in HPC environments, addressing the challenge of real-time interaction limitations. The framework allows workflows to pause for human input at defined checkpoints without interrupting compute jobs, thereby improving resource utilization and enabling non-blocking supervision. The authors demonstrate the framework's effectiveness in model training on HPC systems like MareNostrum 5, showcasing its portability, efficiency, and enhanced oversight in operational AI workflows.
Unlock human-in-the-loop AI workflows on HPC systems without sacrificing compute efficiency using this new asynchronous collaboration framework.
Human involvement is critical in training and deploying AI systems in high-stakes defence and security contexts. However, real-time interaction is impractical in HPC environments due to compute intensity and resource constraints. We present a workflow framework that enables asynchronous human-AI collaboration across hybrid infrastructures, including HPC clusters, local machines, and cloud platforms. Workflows can pause at defined checkpoints for human input without halting underlying compute jobs, preventing idle resources and enabling non-blocking supervision. The framework supports interaction with SLURM-based scheduling, containerized and native tasks, and is customized for scenarios requiring human judgment and adaptability. We demonstrate its application in model training on systems like MareNostrum 5, highlighting benefits in portability, efficiency, and oversight in operational AI workflows.