CAREER: Towards Transformorphic Computing

Software systems are behind almost everything people do today, from science and medicine to schoolwork and communication. Unfortunately, their vast complexity goes far beyond the understanding of any single person or team, making even small improvements in the face of new tasks, environments, or users feasible only for large organizations employing teams of experts. This project aims to develop transformation techniques for automatically improving the performance, security, and reliability of existing and widely deployed systems—allowing anyone to improve complex systems today, rather than waiting years for them to be improved or rebuilt by experts. The proposed project will develop techniques for automated, incremental, and tunable whole-systems transformations. These techniques will be centered around fracture, piecewise transformation, and careful reassembly of entire systems and their polyglot components; domain-specific configurability, specialized inference, and goal-specific regeneration; and runtime support for manipulating effects and confirming the observational equivalence between the original and transformed systems. In parallel, the project will develop educational programs that prepare students to operate on and improve existing, real-world systems. Combined, the project will yield fundamental advances in computing systems, usable open-source software, and a thorough and accessible organization of knowledge on transformorphic computing—helping establish a new research community. By lowering barriers to improving, morphing, and extending existing and widely deployed software systems, the project will benefit scientists, scholars, companies, entrepreneurs, and the general public. It will enable everyday developers and domain experts to adapt and improve software systems with less manual effort and more automation, making advanced capabilities available today—not in a few years or decades. Because the project targets pervasively used environments, its techniques and tools can deliver performance, security, and reliability benefits across the national workforce, industry, government, and open-source communities worldwide. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.