MIT’s Plasma Science and Fusion Centre (PSFC) and its collaborators aim to accelerate the development of fusion energy as a clean energy source. The U.S. Department of Energy (DoE) has funded this three-year project, which focuses on integrating fusion data into a system accessible to AI-powered tools. The project aims to pilot integrating fusion data into a system compatible with AI-driven tools. The researchers intend to construct an all-encompassing fusion data platform.
This platform’s components have the potential to offer unprecedented accessibility to researchers, with a specific emphasis on underrepresented students. The initiative seeks to promote diverse engagement in fusion and data science, both within the academic sphere and the professional arena, through outreach programmes spearheaded by the project’s co-investigators, four of whom are women out of a total of five.
Presently, there exist nearly 50 publicly accessible experimental magnetic confinement fusion devices. Nevertheless, accessing historical and current data from these devices poses challenges. Some fusion databases mandate users to agree to specific terms, and data organisation varies widely across these sources.
Furthermore, using machine learning, which constitutes a specialised subset of artificial intelligence tools, for data analysis and facilitating groundbreaking scientific discoveries can often be arduous. It is primarily attributed to the significant requirement for extensive data restructuring and preprocessing before machine learning algorithms can effectively harness it.
Consequently, this challenge results in a discernible reduction in the number of scientists actively participating in fusion research, a notable increase in the barriers to making important discoveries within the field, and a considerable bottleneck in realising the full potential of artificial intelligence in expediting advancements and breakthroughs in fusion research. It underscores the pressing need for innovative solutions that streamline data preparation and analysis processes to fully leverage the power of AI in the pursuit of fusion energy.
The researchers have set their sights on addressing and surmounting the various barriers that have historically impeded the active participation of women and marginalised groups in the field. Their comprehensive strategy encompasses enhancing overall access to fusion data and includes a visionary initiative: establishing a subsidised summer school programme.
This educational endeavour is meticulously designed to concentrate on the intersection of fusion science and machine learning, two domains at the forefront of scientific innovation. Over the next three years, this groundbreaking summer school will be hosted at the prestigious institution William & Mary.
By pursuing this approach, the research team endeavours to level the playing field, ensuring that individuals from underrepresented backgrounds have equitable access to knowledge and opportunities in fusion research, aiming to foster the development of expertise in fusion science and the cutting-edge applications of machine learning. This initiative encourages diversity and inclusivity within the fusion research community, empowering a new generation of scientists and engineers from diverse backgrounds.
Through this innovative educational programme and their broader efforts to democratise access to fusion data, the researchers aspire to break down longstanding barriers and usher in a future where diverse voices and perspectives contribute to transformative advancements in fusion energy and beyond. This holistic approach aligns with the vision of fostering a more inclusive and dynamic scientific community poised to tackle the complex challenges of the 21st century.
