Experimental science areas are increasingly adopting advanced computing techniques, across a spectrum of computing needs and levels of integration. Across the spectrum from established institutional computing solutions to highly customized approaches, there is a great opportunity for idea sharing and collaboration. While diverse scientific areas from experimental physics, observational astronomy, or lab-bench biomedical research often use myriad approaches to computing and automation, there are many common aspects to managing the intersection of experiment and computation.
ACX encourages participation from computing research to advance discovery applied to diverse domains, including but not limited to: materials science, biological and life science, physics, chemistry, environmental science, geophysics, and planetary science.
Advances in machine learning (ML) and artificial intelligence (AI) capabilities are on the verge of causing major transformations in the experimental science process. These changes are driven by improvements in the numerical engines that power ML/AI capabilities, as well as advances in simulations and digital twins used as references by ML/AI services. Agentic AI is transforming experimental science by autonomously designing, executing, and adapting experiments in real time, accelerating discovery far beyond traditional workflows. Additionally, progress in automated techniques that integrate advanced computing resources with data management and experiment control systems supports these transformations.
There is much work to do. Deeper automation in experimental science will hasten the speed of discovery, but many challenges remain. Automated services that are sufficiently generalized to manage all aspects of the scientific work cycle are not a reality, due to challenges in the complexities of collecting physical data, managing the data lifecycle, actuating computation, and interacting with human users. Researchers and developers working on many aspects of these problems are invited to participate, see the Topics of Interest below.
We envision the ACX Workshop at SupercomputingAsia to be an exciting forum for the exchange of ideas around this topic, from the presentation of inspiring results in “big science” endeavors, to the “day one” creativity that it takes to bring automation into emerging and potentially transformational experimental approaches.
Topics of Interest
New algorithms and approaches for science-focused and science-informed AI
Agentic AI and new interfaces to and uses of large generalized models
Systems to integrate rich experimental data with analysis engines
Automation systems that co-manage computer systems and data collection systems
Applications that combine human and machine insight
Programming, parallelization, and workflow systems for scientific data analysis
Management of scientific data/metadata/provenance
Simulation calibration, data assimilation, and automatic validation of observed data via digital twins or related approaches
Coupling large-scale experimental and observational facilities with large-scale computing systems
Emerging edge technologies for real-time data-driven discovery
Autonomous experimentation driven by HPC