arXiv subjects

Xiong Zhou

Publications and source records attributed to Xiong Zhou.

27 records · Page 2Linked to original sources

arXiv2021

Room temperature coherent manipulation of single-spin qubits in silicon carbide with a high readout contrast

Spin defects in silicon carbide (SiC) with mature wafer-scale fabrication and micro/nano-processing technologies have recently drawn considerable attention. Although room temperature single-spin manipulation of colour centres in SiC has been demonstrated, the typically detected contrast is less than 2%, and the photon count rate is also low. Here, we present the coherent manipulation of single divacancy spins in 4H-SiC with a high readout contrast (-30%) and a high photon count rate (150 kilo counts per second) under ambient conditions, which are competitive with the nitrogen-vacancy (NV) centres in diamond. Coupling between a single defect spin and a nearby nuclear spin is also observed. We further provide a theoretical explanation for the high readout contrast by analysing the defect levels and decay paths. Since the high readout contrast is of utmost importance in many applications of quantum technologies, this work might open a new territory for SiC-based quantum devices with many advanced properties of the host material.

cond-mat.mes-hall↗

arXiv2021

Asymmetric Loss Functions for Learning with Noisy Labels

Robust loss functions are essential for training deep neural networks with better generalization power in the presence of noisy labels. Symmetric loss functions are confirmed to be robust to label noise. However, the symmetric condition is overly restrictive. In this work, we propose a new class of loss functions, namely \textit{asymmetric loss functions}, which are robust to learning with noisy labels for various types of noise. We investigate general theoretical properties of asymmetric loss functions, including classification calibration, excess risk bound, and noise tolerance. Meanwhile, we introduce the asymmetry ratio to measure the asymmetry of a loss function. The empirical results show that a higher ratio would provide better noise tolerance. Moreover, we modify several commonly-used loss functions and establish the necessary and sufficient conditions for them to be asymmetric. Experimental results on benchmark datasets demonstrate that asymmetric loss functions can outperform state-of-the-art methods. The code is available at \href{https://github.com/hitcszx/ALFs}{https://github.com/hitcszx/ALFs}

cs.LG↗

arXiv2020

Advances in Deep Learning for Hyperspectral Image Analysis--Addressing Challenges Arising in Practical Imaging Scenarios

Deep neural networks have proven to be very effective for computer vision tasks, such as image classification, object detection, and semantic segmentation -- these are primarily applied to color imagery and video. In recent years, there has been an emergence of deep learning algorithms being applied to hyperspectral and multispectral imagery for remote sensing and biomedicine tasks. These multi-channel images come with their own unique set of challenges that must be addressed for effective image analysis. Challenges include limited ground truth (annotation is expensive and extensive labeling is often not feasible), and high dimensional nature of the data (each pixel is represented by hundreds of spectral bands), despite being presented by a large amount of unlabeled data and the potential to leverage multiple sensors/sources that observe the same scene. In this chapter, we will review recent advances in the community that leverage deep learning for robust hyperspectral image analysis despite these unique challenges -- specifically, we will review unsupervised, semi-supervised and active learning approaches to image analysis, as well as transfer learning approaches for multi-source (e.g. multi-sensor, or multi-temporal) image analysis.

cs.CV↗

arXiv2020

Coherent control of nitrogen-vacancy center spins in silicon carbide at room temperature

Solid-state color centers with manipulatable spin qubits and telecom-ranged fluorescence are ideal platforms for quantum communications and distributed quantum computations. In this work, we coherently control the nitrogen-vacancy (NV) center spins in silicon carbide at room temperature, in which telecom-wavelength emission is detected. We increase the NV concentration six-fold through optimization of implantation conditions. Hence, coherent control of NV center spins is achieved at room temperature and the coherence time T2 can be reached to around 17.1 μs. Furthermore, investigation of fluorescence properties of single NV centers shows that they are room temperature photostable single photon sources at telecom range. Taking advantages of technologically mature materials, the experiment demonstrates that the NV centers in silicon carbide are promising platforms for large-scale integrated quantum photonics and long-distance quantum networks.

cond-mat.mes-hall↗

arXiv2020

Out-of-the-box channel pruned networks

In the last decade convolutional neural networks have become gargantuan. Pre-trained models, when used as initializers are able to fine-tune ever larger networks on small datasets. Consequently, not all the convolutional features that these fine-tuned models detect are requisite for the end-task. Several works of channel pruning have been proposed to prune away compute and memory from models that were trained already. Typically, these involve policies that decide which and how many channels to remove from each layer leading to channel-wise and/or layer-wise pruning profiles, respectively. In this paper, we conduct several baseline experiments and establish that profiles from random channel-wise pruning policies are as good as metric-based ones. We also establish that there may exist profiles from some layer-wise pruning policies that are measurably better than common baselines. We then demonstrate that the top layer-wise pruning profiles found using an exhaustive random search from one datatset are also among the top profiles for other datasets. This implies that we could identify out-of-the-box layer-wise pruning profiles using benchmark datasets and use these directly for new datasets. Furthermore, we develop a Reinforcement Learning (RL) policy-based search algorithm with a direct objective of finding transferable layer-wise pruning profiles using many models for the same architecture. We use a novel reward formulation that drives this RL search towards an expected compression while maximizing accuracy. Our results show that our transferred RL-based profiles are as good or better than best profiles found on the original dataset via exhaustive search. We then demonstrate that if we found the profiles using a mid-sized dataset such as Cifar10/100, we are able to transfer them to even a large dataset such as Imagenet.

cs.LG↗

arXiv2019

FineText: Text Classification via Attention-based Language Model Fine-tuning

Training deep neural networks from scratch on natural language processing (NLP) tasks requires significant amount of manually labeled text corpus and substantial time to converge, which usually cannot be satisfied by the customers. In this paper, we aim to develop an effective transfer learning algorithm by fine-tuning a pre-trained language model. The goal is to provide expressive and convenient-to-use feature extractors for downstream NLP tasks, and achieve improvement in terms of accuracy, data efficiency, and generalization to new domains. Therefore, we propose an attention-based fine-tuning algorithm that automatically selects relevant contextualized features from the pre-trained language model and uses those features on downstream text classification tasks. We test our methods on six widely-used benchmarking datasets, and achieve new state-of-the-art performance on all of them. Moreover, we then introduce an alternative multi-task learning approach, which is an end-to-end algorithm given the pre-trained model. By doing multi-task learning, one can largely reduce the total training time by trading off some classification accuracy.

cs.CL↗

arXiv2019

$d$-SNE: Domain Adaptation using Stochastic Neighborhood Embedding

Deep neural networks often require copious amount of labeled-data to train their scads of parameters. Training larger and deeper networks is hard without appropriate regularization, particularly while using a small dataset. Laterally, collecting well-annotated data is expensive, time-consuming and often infeasible. A popular way to regularize these networks is to simply train the network with more data from an alternate representative dataset. This can lead to adverse effects if the statistics of the representative dataset are dissimilar to our target. This predicament is due to the problem of domain shift. Data from a shifted domain might not produce bespoke features when a feature extractor from the representative domain is used. In this paper, we propose a new technique ($d$-SNE) of domain adaptation that cleverly uses stochastic neighborhood embedding techniques and a novel modified-Hausdorff distance. The proposed technique is learnable end-to-end and is therefore, ideally suited to train neural networks. Extensive experiments demonstrate that $d$-SNE outperforms the current states-of-the-art and is robust to the variances in different datasets, even in the one-shot and semi-supervised learning settings. $d$-SNE also demonstrates the ability to generalize to multiple domains concurrently.

cs.CV↗

arXiv2018

On-demand generation of shallow silicon vacancy in silicon carbide

Defects in silicon carbide have been explored as promising spin systems in quantum technologies. However, for practical quantum metrology and quantum communication, it is critical to achieve the on-demand shallow spin-defect generation. In this work, we present the generation and characterization of shallow silicon vacancies in silicon carbide by using different implanted ions and annealing conditions. The conversion efficiency of silicon vacancy of helium ions is shown to be higher than that by carbon and hydrogen ions in a wide implanted fluence range. Furthermore, after optimizing annealing conditions, the conversion efficiency can be increased more than 2 times. Due to the high density of the generated ensemble defects, the sensitivity to sense a static magnetic field can be research as high as , which is about 15 times higher than previous results. By carefully optimizing implanted conditions, we further show that a single silicon vacancy array can be generated with about 80 % conversion efficiency, which reaches the highest conversion yield in solid state systems. The results pave the way for using on-demand generated shallow silicon vacancy for quantum information processing and quantum photonics.

quant-ph↗

arXiv2011

Creating and modulating electronic states on noble metal surfaces: ultrathin Ag islands on Si(111)-7$\times$7 as a prototype

Various-thickness Ag islands were prepared on Si(111)-7$\times$7 using the one-step deposition at a high substrate temperature. An electronic state centered at -0.40$\sim$-0.15eV versus E$_{Fermi}$, detectable on the surface of the Ag islands thinner than 9 layers, was created by the electronic hybridization between Ag and Si at the Ag-Si interface. Scanning tunneling microscopy/spectroscopy and density functional theory revealed that the thickness of Ag islands determined the strength of the hybridization, leading to a modulation to the energy and intensity of the state on the surface.

cond-mat.mtrl-sci↗