Source author record

Ye Zhou

Ye Zhou appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

cond-mat.soft Computer Vision eess.SY Machine Learning physics.plasm-ph Systems and Control

Catalog footprint

What is connected

5works

6topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2022arXiv

Role of self-generated magnetic fields in the inertial fusion ignition threshold

Magnetic fields spontaneously grow at unstable interfaces around hot-spot asymmetries during inertial confinement fusion implosions. Although difficult to measure, theoretical considerations and numerical simulations predict field strengths exceeding 5kT in current national ignition facility experiments. Magnetic confinement of electrons then reduces the hot-spot heat loss by >5%. We demonstrate this via magnetic post-processing of two-dimensional xRAGE hydrodynamic simulation data at bang time. We then derive a model for the self-magnetization, finding that it varies with the square of the stagnation temperature and inversely with the areal density. The self-magnetized Lawson analysis then gives a slightly reduced ignition threshold. Time dependent hot-spot energy balance models corroborate this finding, with the magnetic field quadrupling the fusion yield for near threshold parameters. The inclusion of magnetized multi-dimensional fluid instabilities could further alter the ignition threshold, and will be the subject of future work.

preprint2021arXiv

Rethinking Soft Labels for Knowledge Distillation: A Bias-Variance Tradeoff Perspective

Knowledge distillation is an effective approach to leverage a well-trained network or an ensemble of them, named as the teacher, to guide the training of a student network. The outputs from the teacher network are used as soft labels for supervising the training of a new network. Recent studies \citep{muller2019does,yuan2020revisiting} revealed an intriguing property of the soft labels that making labels soft serves as a good regularization to the student network. From the perspective of statistical learning, regularization aims to reduce the variance, however how bias and variance change is not clear for training with soft labels. In this paper, we investigate the bias-variance tradeoff brought by distillation with soft labels. Specifically, we observe that during training the bias-variance tradeoff varies sample-wisely. Further, under the same distillation temperature setting, we observe that the distillation performance is negatively associated with the number of some specific samples, which are named as regularization samples since these samples lead to bias increasing and variance decreasing. Nevertheless, we empirically find that completely filtering out regularization samples also deteriorates distillation performance. Our discoveries inspired us to propose the novel weighted soft labels to help the network adaptively handle the sample-wise bias-variance tradeoff. Experiments on standard evaluation benchmarks validate the effectiveness of our method. Our code is available at \url{https://github.com/bellymonster/Weighted-Soft-Label-Distillation}.

preprint2020arXiv

Active Gate Drive with Gate-Drain Discharge Compensation for Voltage Balancing in Series-Connected SiC MOSFETs

Imbalanced voltage sharing during the turn-off transient is a challenge for series-connected silicon carbide (SiC) MOSFET application. This article first discusses the influence of the gate-drain discharge deviation on the voltage imbalance ratio, and its primary causes are also presented and verified by LTspice simulation. Accordingly, a novel active gate drive, which aims to compensate the discharge difference between devices connected in series, is proposed and analyzed. By only using the original output of the driving IC, the proposed gate drive is realized by implementing an auxiliary circuit on the existing commercial gate drive. Therefore, unlike other active gate drives for balancing control, no extra isolations for power/signal are needed, and the number of the devices in series is unlimited. The auxiliary circuit includes three sub-circuits as a high-bandwidth current sink for regulating switching performance, a relative low-frequency but reliable sampling and control circuit for closed-loop control, and a trigger combining the former and the latter. The operational principle and the design guideline for each part are presented in detail. Experimental results validate the performance of the proposed gate drive and its voltage balancing control algorithm.

preprint2020arXiv

Structural transformations in tetravalent nematic shells induced by a magnetic field

The role of applied fields on the structure of liquid crystals confined to shell geometries has been studied in past theoretical work, providing strategies to produce liquid crystal shells with controlled defect structure or valence. However, the predictions of such studies have not been experimentally explored yet. In this work, we study the structural transformations undergone by tetravalent nematic liquid crystal shells under a strong uniform magnetic field, using both experiments and simulations. We consider two different cases in terms of shell geometry and initial defect symmetry: i) homogeneous shells with four s = +1/2 defects in a tetrahedral arrangement, and ii) inhomogeneous shells with four s = +1/2 defects localized in their thinner parts. Consistently with previous theoretical results, we observe that the initial defect structure evolves into a bipolar one, in a process where the defects migrate towards the poles. Interestingly, we find that the defect trajectories and dynamics are controlled by curvature walls that connect the defects by pairs. Based on the angle between Bs, the local projection of the magnetic field on the shell surface, and n+ 1/2 , a vector describing the defect orientations, we are able to predict the nature and shape of those inversion walls, and therefore, the trajectory and dynamics of the defects. This rule, based on symmetry arguments, is consistent with both experiments and simulations and applies for shells that are either homogeneous or inhomogeneous in thickness. By modifying the angle between Bs and n+1/2, we are able to induce, in controlled way, complex routes towards the final bipolar state. In the case of inhomogeneous shells, the specific symmetry of the shell allowed us to observe a hybrid splay-bend Helfrich wall for the first time.

preprint2016arXiv

Mesoscale Structure of Chiral Nematic Shells

There is considerable interest in understanding and controlling topological defects in nematic liquid crystals (LCs). Confinement, in the form of droplets, has been particularly effective in that regard. Here, we employ the Landau-de Gennes method to explore the geometrical frustration of nematic order in shell geometries, and focus on chiral materials. By varying the chirality and thickness in uniform shells, we construct a phase diagram that includes tetravalent structures, bipolar structures (BS), bent structures and radial spherical structures (RSS). It is found that, in uniform shells, the BS-to-RSS structural transition, in response to both chirality and shell geometry, is accompanied by an abrupt change of defect positions, implying a potential use for chiral nematic shells as sensors. Moreover, we investigate thickness heterogeneity in shells and demonstrate that non-chiral and chiral nematic shells exhibit distinct equilibrium positions of their inner core that are governed by shell chirality c.