Doan Nguyen | Searcharxiv

arXiv2023

Nondestructive testing of high strength conductors for high field pulsed magnets

High field pulsed magnets at the NHMFL use high strength conductor wires up to 90% of their ultimate tensile strength. Therefore it is very important to ensure that the wires are free of flaws. It is known that in the conductors cold drawing process, internal chevron crack could occur due to unsuitable drawing die schedule or inadequate lubrication. These internal cracks occurs infrequently along the wire, so tensile tests of short samples cut from the ends of a long length conductor often miss the problem. In addition, small inclusions on the wire surface can compromise wires fatigue properties. In this paper, we present results of our non-destructive testing (NDT) inspection of Glidcop AL60 wires using eddy current testing (ECT), ultrasonic testing (UT) and x-ray radiography (2D and 3D). Chevron cracks were found in some AL60 conductors by all three NDT techniques. Surface inclusions were found by ECT. We have developed a long length ECT wire inspection capability.

cond-mat.mtrl-sci↗

arXiv2022

Split Learning without Local Weight Sharing to Enhance Client-side Data Privacy

Split learning (SL) aims to protect user data privacy by distributing deep models between client-server and keeping private data locally. In SL training with multiple clients, the local model weights are shared among the clients for local model update. This paper first reveals data privacy leakage exacerbated from local weight sharing among the clients in SL through model inversion attacks. Then, to reduce the data privacy leakage issue, we propose and analyze privacy-enhanced SL (P-SL) (or SL without local weight sharing). We further propose parallelized P-SL to expedite the training process by duplicating multiple server-side model instances without compromising accuracy. Finally, we explore P-SL with late participating clients and devise a server-side cache-based training method to address the forgetting phenomenon in SL when late clients join. Experimental results demonstrate that P-SL helps reduce up to 50% of client-side data leakage, which essentially achieves a better privacy-accuracy trade-off than the current trend by using differential privacy mechanisms. Moreover, P-SL and its cache-based version achieve comparable accuracy to baseline SL under various data distributions, while cost less computation and communication. Additionally, caching-based training in P-SL mitigates the negative effect of forgetting, stabilizes the learning, and enables practical and low-complexity training in a dynamic environment with late-arriving clients.

cs.CR↗