Published in Machine Learning: Science and Technology, 2026
Unsupervised flow matching learns compact, scale-aware representations of cosmological simulations for reconstruction, synthetic data generation, and parameter estimation.
Published in Nature Communications, 2024
A reconfigurable probabilistic-bit Ising-machine architecture that emulates all-to-all connectivity while retaining highly parallel sampling for hard combinatorial optimization.
Published in NeurIPS 2025 Workshop on AI for Science, 2025
Graph neural network surrogates that reproduce LIGO interferometer simulations up to 815 times faster, supporting more efficient instrumentation design.
Published in ML4Astro Workshop, Co-located with ICML 2025, 2025
A flow-matching framework that compresses cosmological field simulations into compact, interpretable representations for reconstruction, generation, and parameter inference.
Published in 51st ACM/IEEE International Symposium on Computer Architecture (ISCA), 2024
A systems architecture that physically moves SSD-resident datasets through data centres, reducing data-movement time and energy for large-scale workloads.