I am a computer scientist with a research focus on evolutionary algorithms (EA) within a High Performance Computing (HPC) context. EAs are well positioned to exploit HPC platforms because of their scalable and embarrassingly parallel nature, and ORNL is ideal for research in this area due to ready access to HPC platforms, such as Summit.
Asynchronous steady-state EAs (ASEA) are typically used for HPC-related work, and some of my most recent work entailed using ASEAs to optimize deep-learner (DL) hyper-parameters and architectures on Summit. However, we are still getting an understanding of the unique dynamics of ASEAs to scale on very large, Summit-sized systems. I am actively engaged in improving our understanding of ASEAs on such platforms to provide useful guidelines for practitioners that wish to use ASEAs to solve real-world problems.
July 2009, Best Graduate Student Workshop Paper at GECCO, Coletti, Mark. "Learnable evolution model performance impaired by binary tournament survival selection." In Proceedings of the 11th Annual Conference Companion on Genetic and Evolutionary Computation Conference: Late Breaking Papers, pp. 2717-2720. 2009.
November 2021, UKCI, Best paper, Scott, E.O. et al. (2022). Avoiding Excess Computation in Asynchronous Evolutionary Algorithms. In: Jansen, T., Jensen, R., Mac Parthaláin, N., Lin, CM. (eds) Advances in Computational Intelligence Systems. UKCI 2021. Advances in Intelligent Systems and Computing, vol 1409. Springer, Cham. https://doi.org/10.1007/978-3-030-87094-2_7
2022, R&D 100 Award, Gremlin: Discovering Weaknesses in Artificial Intelligence
I use the Summit supercomputer platform for my evolutionary computation / deep learner research.