arXiv subjects

Qian Wang

Publications and source records attributed to Qian Wang.

At least 433 records · Page 24Linked to original sources

arXiv2021

Black-box Adversarial Attacks on Commercial Speech Platforms with Minimal Information

Adversarial attacks against commercial black-box speech platforms, including cloud speech APIs and voice control devices, have received little attention until recent years. The current "black-box" attacks all heavily rely on the knowledge of prediction/confidence scores to craft effective adversarial examples, which can be intuitively defended by service providers without returning these messages. In this paper, we propose two novel adversarial attacks in more practical and rigorous scenarios. For commercial cloud speech APIs, we propose Occam, a decision-only black-box adversarial attack, where only final decisions are available to the adversary. In Occam, we formulate the decision-only AE generation as a discontinuous large-scale global optimization problem, and solve it by adaptively decomposing this complicated problem into a set of sub-problems and cooperatively optimizing each one. Our Occam is a one-size-fits-all approach, which achieves 100% success rates of attacks with an average SNR of 14.23dB, on a wide range of popular speech and speaker recognition APIs, including Google, Alibaba, Microsoft, Tencent, iFlytek, and Jingdong, outperforming the state-of-the-art black-box attacks. For commercial voice control devices, we propose NI-Occam, the first non-interactive physical adversarial attack, where the adversary does not need to query the oracle and has no access to its internal information and training data. We combine adversarial attacks with model inversion attacks, and thus generate the physically-effective audio AEs with high transferability without any interaction with target devices. Our experimental results show that NI-Occam can successfully fool Apple Siri, Microsoft Cortana, Google Assistant, iFlytek and Amazon Echo with an average SRoA of 52% and SNR of 9.65dB, shedding light on non-interactive physical attacks against voice control devices.

cs.CR↗

arXiv2021

Study of heavy quark conserving weak decays in the quark model

Within the framework of a constituent quark model, we investigate the heavy quark conserving weak decays of $Ξ_c$ and $Ξ_b$. This process involves elementary transitions of $us\to du$ or $s\to ud \bar u$ with the heavy quark as a spectator, while for the charmed baryon in the initial state it can also occur via $cs\to dc$ to conserve the heavy quark flavor in the weak decays. It shows that the heavy quark symmetry (HQS) plays an essential role in the processes where the heavy quark acts as a spectator. For the $Ξ_c$ decays, the dominance of the pole terms in the parity-conserving transitions is evident. This can naturally explain the sizeable branching ratio for $Ξ_c^0\toΛ_cπ^-$ as recently reported by LHCb. The same mechanism predicts a large branching ratio for $Ξ_c^+\toΛ_cπ^0$. In contrast, the decays of $Ξ_b$ will only go through the parity-violating processes because of the HQS. In addition, the branching ratios of the $Ξ_c^+\to Λ_c π^0$ and $Ξ_b^0\to Λ_b π^0$ processes are also predicted for the further experimental measurement.

hep-ph↗

arXiv2021

On Novel Peer Review System for Academic Journals: Experimental Study Based on Social Computing

For improving the performance and effectiveness of peer review, a novel review system is proposed, based on analysis of peer review process for academic journals under a parallel model built via Monte Carlo method. The model can simulate the review, application and acceptance activities of the review systems, in a distributed manner. According to simulation experiments on two distinct review systems respectively, significant advantages manifest for the novel one.

cs.DL↗

arXiv2021

Unsupervised Landmark Detection Based Spatiotemporal Motion Estimation for 4D Dynamic Medical Images

Motion estimation is a fundamental step in dynamic medical image processing for the assessment of target organ anatomy and function. However, existing image-based motion estimation methods, which optimize the motion field by evaluating the local image similarity, are prone to produce implausible estimation, especially in the presence of large motion. In this study, we provide a novel motion estimation framework of Dense-Sparse-Dense (DSD), which comprises two stages. In the first stage, we process the raw dense image to extract sparse landmarks to represent the target organ anatomical topology and discard the redundant information that is unnecessary for motion estimation. For this purpose, we introduce an unsupervised 3D landmark detection network to extract spatially sparse but representative landmarks for the target organ motion estimation. In the second stage, we derive the sparse motion displacement from the extracted sparse landmarks of two images of different time points. Then, we present a motion reconstruction network to construct the motion field by projecting the sparse landmarks displacement back into the dense image domain. Furthermore, we employ the estimated motion field from our two-stage DSD framework as initialization and boost the motion estimation quality in light-weight yet effective iterative optimization. We evaluate our method on two dynamic medical imaging tasks to model cardiac motion and lung respiratory motion, respectively. Our method has produced superior motion estimation accuracy compared to existing comparative methods. Besides, the extensive experimental results demonstrate that our solution can extract well representative anatomical landmarks without any requirement of manual annotation. Our code is publicly available online.

eess.IV↗

arXiv2021

Reconstructing Training Data from Diverse ML Models by Ensemble Inversion

Model Inversion (MI), in which an adversary abuses access to a trained Machine Learning (ML) model attempting to infer sensitive information about its original training data, has attracted increasing research attention. During MI, the trained model under attack (MUA) is usually frozen and used to guide the training of a generator, such as a Generative Adversarial Network (GAN), to reconstruct the distribution of the original training data of that model. This might cause leakage of original training samples, and if successful, the privacy of dataset subjects will be at risk if the training data contains Personally Identifiable Information (PII). Therefore, an in-depth investigation of the potentials of MI techniques is crucial for the development of corresponding defense techniques. High-quality reconstruction of training data based on a single model is challenging. However, existing MI literature does not explore targeting multiple models jointly, which may provide additional information and diverse perspectives to the adversary. We propose the ensemble inversion technique that estimates the distribution of original training data by training a generator constrained by an ensemble (or set) of trained models with shared subjects or entities. This technique leads to noticeable improvements of the quality of the generated samples with distinguishable features of the dataset entities compared to MI of a single ML model. We achieve high quality results without any dataset and show how utilizing an auxiliary dataset that's similar to the presumed training data improves the results. The impact of model diversity in the ensemble is thoroughly investigated and additional constraints are utilized to encourage sharp predictions and high activations for the reconstructed samples, leading to more accurate reconstruction of training images.

cs.LG↗

arXiv2021

Noble-Metal-Assisted Fast Interfacial Oxygen Migration with Topotactic Phase Transition in Perovskite Oxides

Transition-metal perovskite oxides constitute a series of functional material systems for electronics, catalysis and energy-conversion processes, in which oxygen migration and evolution play a key role. However, the stable metal-oxygen (M-O) bond forms large energy barrier inhibiting ion diffusion. Therefore, seeking efficient and facile approaches to accelerate oxygen kinetics has become a significant issue. Here, the interaction (interfacial charge transfer and cooperative bonding) between noble metal (Pt, Ag) and perovskites oxide (SrCoO3-delta) is employed to weaken (M-O) bond and decrease the energy barrier of oxygen migration. Noble metal layers serving as oxygen pumps can continuously extract oxygen from oxide films to atmosphere. The temperature of topotactic phase transition from perovskite (SrCoO3) to brownmillerite (SrCoO2.5) is remarkably lowered from 200 K to room temperature. Furthermore, this approach can also be applied to SrFeO3 for similar topotactic phase reduction by moderate thermal activation. Our finding paves a promising and general pathway to achieve fast oxygen migration in perovskite oxides, with important application prospect in low-temperature electrodes and high-activity catalysis interface.

cond-mat.mtrl-sci↗

arXiv2021

Interface-enhanced ferromagnetism with long-distance effect in van der Waals semiconductor

Ferromagnetic semiconductors discovered in two-dimensional (2D) materials open an avenue for highly integrated and multifunctional spintronics. The Curie temperature (TC) of existed 2D ferromagnetic semiconductors is extremely low and the modulation effect of their magnetism is limited compared with their 2D metallic counterparts. The interfacial effect was found to effectively manipulate the three-dimensional magnetism, providing a unique opportunity for tailoring the 2D magnetism. Here we demonstrate that the TC of 2D ferromagnetic semiconductor Cr2Ge2Te6 can be enhanced by 130% (from ~65 K to above 150 K) when adjacent to a tungsten layer. The interfacial W-Te bonding contributes to the TC enhancement with a strong perpendicular magnetic anisotropy (PMA), guaranteeing an efficient magnetization switching by the spin-orbit torque with a low current density at 150 K. Distinct from the rapid attenuation in conventional magnets, the interfacial effect exhibits a weak dependence on Cr2Ge2Te6 thickness and a long-distance effect (more than 10 nanometers) due to the weak interlayer coupling inherent to 2D magnets. Our work not only reveals a unique interfacial behavior in 2D materials, but also advances the process towards practical 2D spintronics.

cond-mat.mtrl-sci↗

arXiv2021

Progressively Select and Reject Pseudo-labelled Samples for Open-Set Domain Adaptation

Domain adaptation solves image classification problems in the target domain by taking advantage of the labelled source data and unlabelled target data. Usually, the source and target domains share the same set of classes. As a special case, Open-Set Domain Adaptation (OSDA) assumes there exist additional classes in the target domain but not present in the source domain. To solve such a domain adaptation problem, our proposed method learns discriminative common subspaces for the source and target domains using a novel Open-Set Locality Preserving Projection (OSLPP) algorithm. The source and target domain data are aligned in the learned common spaces class-wisely. To handle the open-set classification problem, our method progressively selects target samples to be pseudo-labelled as known classes and rejects the outliers if they are detected as from unknown classes. The common subspace learning algorithm OSLPP simultaneously aligns the labelled source data and pseudo-labelled target data from known classes and pushes the rejected target data away from the known classes. The common subspace learning and the pseudo-labelled sample selection/rejection facilitate each other in an iterative learning framework and achieves state-of-the-art performance on benchmark datasets Office-31 and Office-Home with the average HOS of 87.4% and 67.0% respectively.

cs.CV↗

arXiv2021

Multifractality in quasienergy space of coherent states as a signature of quantum chaos

We present the multifractal analysis of coherent states in kicked top model by expanding them in the basis of Floquet operator eigenstates. We demonstrate the manifestation of phase space structures in the multifractal properties of coherent states. In the classical limit, the classical dynamical map can be constructed, allowing us to explore the corresponding phase space portraits and to calculate Lyapunov exponent. By tuning the kicking strength, the system undergoes a transition from regularity to chaos. We show that the variation of multifractal dimensions of coherent states with kicking strength is able to capture the structural changes of the phase space. The onset of chaos is clearly identified by the phase space averaged multifractal dimensions, which are well described by random matrix theory in strongly chaotic regime. We further investigate the probability distribution of expansion coefficients, and show that the deviation between the numerical results and the prediction of random matrix theory behaves as a reliable detector of quantum chaos.

quant-ph↗

arXiv2021

Cross-Domain Structure Preserving Projection for Heterogeneous Domain Adaptation

Heterogeneous Domain Adaptation (HDA) addresses the transfer learning problems where data from the source and target domains are of different modalities (e.g., texts and images) or feature dimensions (e.g., features extracted with different methods). It is useful for multi-modal data analysis. Traditional domain adaptation algorithms assume that the representations of source and target samples reside in the same feature space, hence are likely to fail in solving the heterogeneous domain adaptation problem. Contemporary state-of-the-art HDA approaches are usually composed of complex optimization objectives for favourable performance and are therefore computationally expensive and less generalizable. To address these issues, we propose a novel Cross-Domain Structure Preserving Projection (CDSPP) algorithm for HDA. As an extension of the classic LPP to heterogeneous domains, CDSPP aims to learn domain-specific projections to map sample features from source and target domains into a common subspace such that the class consistency is preserved and data distributions are sufficiently aligned. CDSPP is simple and has deterministic solutions by solving a generalized eigenvalue problem. It is naturally suitable for supervised HDA but has also been extended for semi-supervised HDA where the unlabelled target domain samples are available. Extensive experiments have been conducted on commonly used benchmark datasets (i.e. Office-Caltech, Multilingual Reuters Collection, NUS-WIDE-ImageNet) for HDA as well as the Office-Home dataset firstly introduced for HDA by ourselves due to its significantly larger number of classes than the existing ones (65 vs 10, 6 and 8). The experimental results of both supervised and semi-supervised HDA demonstrate the superior performance of our proposed method against contemporary state-of-the-art methods.

cs.LG↗

arXiv2021

Digital Twins based Day-ahead Integrated Energy System Scheduling under Load and Renewable Energy Uncertainties

By constructing digital twins (DT) of an integrated energy system (IES), one can benefit from DT's predictive capabilities to improve coordinations among various energy converters, hence enhancing energy efficiency, cost savings and carbon emission reduction. This paper is motivated by the fact that practical IESs suffer from multiple uncertainty sources, and complicated surrounding environment. To address this problem, a novel DT-based day-ahead scheduling method is proposed. The physical IES is modelled as a multi-vector energy system in its virtual space that interacts with the physical IES to manipulate its operations. A deep neural network is trained to make statistical cost-saving scheduling by learning from both historical forecasting errors and day-ahead forecasts. Case studies of IESs show that the proposed DT-based method is able to reduce the operating cost of IES by 63.5%, comparing to the existing forecast-based scheduling methods. It is also found that both electric vehicles and thermal energy storages play proactive roles in the proposed method, highlighting their importance in future energy system integration and decarbonisation.

cs.IT↗

arXiv2021

Prompt production of the hidden charm pentaquarks in the LHC

Motivated by the observation of the first hidden charm pentaquarks by the LHCb collaboration in 2015 and the updated analysis with an order-of-magnitude larger data set in 2019, we estimate their cross sections for the prompt production as well as their heavy quark spin partners, in the $Σ_c^{(*)}\bar{D}^{(*)}$ hadronic molecular picture, at the center-of-mass energy $7~\mathrm{TeV}$ in the $pp$ collision. Their cross sections are several $\mathrm{nb}$ and we would expect several tens hidden charm pentaquark events in the LHC based on its current integrated luminosity. The cross sections show a sizable deviation of the cross sections for hidden charm pentaquarks with the third isospin component $I_z=+\frac{1}{2}$ ($P_c^+$) from those with $I_z=-\frac{1}{2}$ ($P_c^0$). The cross sections decrease dramatically with the increasing transverse momentum. Our study can also tell where to search for the missing hidden charm pentaquarks. The confirmation of the complete hidden charm pentaquarks in the heavy quark symmetry would further verify their $Σ_c^{(*)}\bar{D}^{(*)}$ molecular interpretation. In addition, the relative strength among these cross sections for pentaquarks can help us to identify the quantum numbers of the $P_c(4440)$ and $P_c(4457)$.

hep-ph↗

arXiv2021

The production of doubly charmed exotic hadrons in heavy ion collisions

Hadron spectroscopy provides direct physical measurements that shed light on the non-perturbative behavior of quantum chromodynamics (QCD). In particular, various exotic hadrons such as the newly observed $T_{cc}^+$ by the LHCb collaboration, offer unique insights on the QCD dynamics in hadron structures. In this letter, we demonstrate how heavy ion collisions can serve as a powerful venue for hadron spectroscopy study of doubly charmed exotic hadrons by virtue of the extremely charm-rich environment created in such collisions. The yields of $T_{cc}^+$ as well as its potential isospin partners are computed within the molecular picture for Pb-Pb collisions at center-of-mass energy $2.76~\mathrm{TeV}$. We find about three-order-of-magnitude enhancement in the production of $T_{cc}^+$ in Pb-Pb collisions as compared with the yield in proton-proton collisions, with a moderately smaller enhancement in the yields of the isospin partners $T_{cc}^0$ and $T_{cc}^{++}$. The $T_{cc}^+$ yield is comparable to that of the $X(3872)$ in the most central collisions while shows a considerably stronger decrease toward peripheral collisions, due to a "threshold" effect of the required double charm quarks for $T_{cc}^+$. Final results for their rapidity and transverse momentum $p_T$ dependence as well as the elliptic flow coefficient are reported and can be tested by future experimental measurements.

hep-ph↗

arXiv2021

DeepCervix: A Deep Learning-based Framework for the Classification of Cervical Cells Using Hybrid Deep Feature Fusion Techniques

Cervical cancer, one of the most common fatal cancers among women, can be prevented by regular screening to detect any precancerous lesions at early stages and treat them. Pap smear test is a widely performed screening technique for early detection of cervical cancer, whereas this manual screening method suffers from high false-positive results because of human errors. To improve the manual screening practice, machine learning (ML) and deep learning (DL) based computer-aided diagnostic (CAD) systems have been investigated widely to classify cervical pap cells. Most of the existing researches require pre-segmented images to obtain good classification results, whereas accurate cervical cell segmentation is challenging because of cell clustering. Some studies rely on handcrafted features, which cannot guarantee the classification stage's optimality. Moreover, DL provides poor performance for a multiclass classification task when there is an uneven distribution of data, which is prevalent in the cervical cell dataset. This investigation has addressed those limitations by proposing DeepCervix, a hybrid deep feature fusion (HDFF) technique based on DL to classify the cervical cells accurately. Our proposed method uses various DL models to capture more potential information to enhance classification performance. Our proposed HDFF method is tested on the publicly available SIPAKMED dataset and compared the performance with base DL models and the LF method. For the SIPAKMED dataset, we have obtained the state-of-the-art classification accuracy of 99.85%, 99.38%, and 99.14% for 2-class, 3-class, and 5-class classification. Moreover, our method is tested on the Herlev dataset and achieves an accuracy of 98.32% for binary class and 90.32% for 7-class classification.

eess.IV↗

arXiv2021

Reducing Magnetic Resonance Image Spacing by Learning Without Ground-Truth

High-quality magnetic resonance (MR) image, i.e., with near isotropic voxel spacing, is desirable in various scenarios of medical image analysis. However, many MR acquisitions use large inter-slice spacing in clinical practice. In this work, we propose a novel deep-learning-based super-resolution algorithm to generate high-resolution (HR) MR images with small slice spacing from low-resolution (LR) inputs of large slice spacing. Notice that most existing deep-learning-based methods need paired LR and HR images to supervise the training, but in clinical scenarios, usually no HR images will be acquired. Therefore, our unique goal herein is to design and train the super-resolution network with no real HR ground-truth. Specifically, two training stages are used in our method. First, HR images of reduced slice spacing are synthesized from real LR images using variational auto-encoder (VAE). Although these synthesized HR images are as realistic as possible, they may still suffer from unexpected morphing induced by VAE, implying that the synthesized HR images cannot be paired with the real LR images in terms of anatomical structure details. In the second stage, we degrade the synthesized HR images to generate corresponding LR images and train a super-resolution network based on these synthesized HR and degraded LR pairs. The underlying mechanism is that such a super-resolution network is less vulnerable to anatomical variability. Experiments on knee MR images successfully demonstrate the effectiveness of our proposed solution to reduce the slice spacing for better rendering.

eess.IV↗

arXiv2021

FederatedNILM: A Distributed and Privacy-preserving Framework for Non-intrusive Load Monitoring based on Federated Deep Learning

Non-intrusive load monitoring (NILM), which usually utilizes machine learning methods and is effective in disaggregating smart meter readings from the household-level into appliance-level consumptions, can help to analyze electricity consumption behaviours of users and enable practical smart energy and smart grid applications. However, smart meters are privately owned and distributed, which make real-world applications of NILM challenging. To this end, this paper develops a distributed and privacy-preserving federated deep learning framework for NILM (FederatedNILM), which combines federated learning with a state-of-the-art deep learning architecture to conduct NILM for the classification of typical states of household appliances. Through extensive comparative experiments, the effectiveness of the proposed FederatedNILM framework is demonstrated.

cs.LG↗

arXiv2021

Electrical peak demand forecasting- A review

The power system is undergoing rapid evolution with the roll-out of advanced metering infrastructure and local energy applications (e.g. electric vehicles) as well as the increasing penetration of intermittent renewable energy at both transmission and distribution level, which characterizes the peak load demand with stronger randomness and less predictability and therefore poses a threat to the power grid security. Since storing large quantities of electricity to satisfy load demand is neither economically nor environmentally friendly, effective peak demand management strategies and reliable peak load forecast methods become essential for optimizing the power system operations. To this end, this paper provides a timely and comprehensive overview of peak load demand forecast methods in the literature. To our best knowledge, this is the first comprehensive review on such topic. In this paper we first give a precise and unified problem definition of peak load demand forecast. Second, 139 papers on peak load forecast methods were systematically reviewed where methods were classified into different stages based on the timeline. Thirdly, a comparative analysis of peak load forecast methods are summarized and different optimizing methods to improve the forecast performance are discussed. The paper ends with a comprehensive summary of the reviewed papers and a discussion of potential future research directions.

eess.SY↗

arXiv2021

Revisiting the nature of the $P_c$ pentaquarks

The nature of the three narrow $P_c$ states, i.e., $P_c(4312)$, $P_c(4440)$ and $P_c(4457)$, is under intense discussion since their discovery from the updated analysis by LHCb. In this work we extend our previous coupled-channel approach [Phys. Rev. Lett. \bf{124}, 072001 (2020)] by including the $Λ_c\bar{D}^{(*)}$ and $η_cp$ as explicitly in addition to the $J/ψp$, as required by unitarity and heavy quark spin symmetry (HQSS). Since inelastic parameters are very badly constrained by the current data, three schemes are considered: (a) scheme I with pure contact interactions and without the $Λ_c\bar D^{(*)}$ interactions, (b) scheme II, where the one-pion exchange is added to scheme I, and (c) scheme III, where the $Λ_c \bar D^{(*)}$ are included in addition. It is shown that to obtain cutoff independent results, OPE in the multichannel system is to be supplemented with $S$-$D$ mixing contact terms. We demonstrate that the experimental data for the $J/ψp$ invariant mass distribution are consistent with the interpretation of the $P_c(4312)$ and $P_c(4440)/P_c(4457)$ as $Σ_c\bar{D}$ and $Σ_c \bar{D}^{*}$ hadronic molecules, respectively, and that the data show clear evidence for a new narrow $P_c(4380)$, as a $Σ_c^*\bar D$ molecule, which should exist as a consequence of HQSS. While two equally good solutions are found in scheme I, only one of these solutions with the quantum numbers of the $P_c(4440)$ and $P_c(4457)$ being $J^P=3/2^-$ and $1/2^-$, respectively, survives the requirement of regulator independence once the OPE is included. Moreover, we predict the line shapes in the elastic and inelastic channels and demonstrate that those related to the $P_c(4440)$ and the $P_c(4457)$ in the corresponding $Σ_c^{(*)}\bar{D}$ and $η_cp$ mass distributions allow one to confirm the quantum numbers given above, once the data are available.

hep-ph↗