Applied Physics Reviews· 2024
Physics-Informed Learning in Artificial Electromagnetic Materials
Deng, Y., Fan, K., Jin, B., Malof, J. M., & Padilla, W. J.
Boston · MTS, KronosAI
Yang
Deng
Charting the space between physics and machine learning.
I build AI-powered simulation tools as an early engineer at KronosAI. Previously, I earned a Ph.D. at Duke studying machine learning for next-generation metamaterials — neural surrogates, inverse design, and active learning at the edge of physics.
Featured · In development
Bayesian optimization,
with an AI copilot.
An interactive notebook for optimization-under-constraints. Three modes — manual, guided, autonomous — each a different amount of trust given to the agent proposing the next experiment.
Manual
Guided
Autonomous
iter · 6 / budget · 24argmax μ + βσ
Selected
publications
Nature Sustainability· 2024equal contribution
Solution-processable bio-inspired smart skin for synergistic solar and radiative heat management
Xie, W., Deng, Y., Liu, Y., et al. (33 authors)
NeurIPS D&B· 2021
Benchmarking data-driven surrogate simulators for artificial electromagnetic materials
Deng, Y., Dong, J., Ren, S., et al.
What I'm thinking about
01
Benchmark design
Automated verification for simulation-based physics discovery.
02
Physics-informed ML
Surrogates, inverse design, active learning — under physical constraints.
03
Foundation models for science
Transformer-based LLMs as high-dimensional scientific regressors.
04
Accelerated optimization
Bayesian methods that make expensive simulations cheaper.
Currently — Member of Technical Staff at KronosAI, building AI-powered simulation tools as an early engineer. Joined as the 2nd engineer; helped bring on the 3rd.