Two Teachers Better Than One: Hardware-Physics Co-Guided Distributed Scientific Machine Learning
Published in Design Automation Conference (DAC), 2026
This paper presents EPIC, a hardware- and physics-co-guided distributed SciML framework for full-waveform inversion. It explores how edge-side encoding and physics-aware decoding can reduce communication latency and energy while preserving physical fidelity.
Recommended citation: Yuchen Yuan, Junhuan Yang, Hao Wan, Yipei Liu, Hanhan Wu, Youzuo Lin and Lei Yang. (2026). "Two Teachers Better Than One: Hardware-Physics Co-Guided Distributed Scientific Machine Learning." Proceedings of the 63rd Design Automation Conference (DAC).