Human-in-the-Loop Machine Learning for Automated Experiments

Description

Human-in-the-loop machine learning (hML) represents a paradigm that bridges the gap between autonomous AI systems and the indispensable intuition and creativity of human operators. In this model, a machine learning (ML) agent autonomously controls experimental workflows, while the human operator monitors and dynamically adjusts the ML agent's strategy. This method addresses the limitations of reinforcement learning in real-world applications by combining the speed and consistency of automation with the adaptive and innovative capabilities of human intelligence. By directly influencing the ML policies and objectives based on real-time observations and outcomes, this system ensures that experiments remain aligned with evolving scientific goals, thereby fostering discovery and optimization in material science, energy, and beyond.

Benefits

• Synergizes human insight with AI's processing power, optimizing experimental outcomes
• Increases efficiency and reproducibility in scientific research and development
• Facilitates the discovery of novel materials and solutions by integrating human creativity into automated processes

Applications and Industries

• Materials science for discovery and optimization of new materials
• Pharmaceutical industry for accelerated drug discovery
• Materials science for developing efficient solar cells and batteries
• Automation and robotics for enhanced decision-making processes

Related ORNL Technologies

• Correlative Multimodal Chemical Imaging via Machine Learning
• Machine Learning on Multimodal Chemical and Whole Slide Imaging Data for Predicting Cancer Labels Directly from Tissue Images
• Pan-Sharpening for Microscopy
• Science-Driven Automated Experiments
• Hypothesis Learning in Automated Experiments

Contact

To learn more about this technology, email partnerships@ornl.gov or call 865-574-1051.