arXiv subjects

Jie Zhang

Publications and source records attributed to Jie Zhang.

At least 919 records · Page 51Linked to original sources

arXiv2020

Attributes Aware Face Generation with Generative Adversarial Networks

Recent studies have shown remarkable success in face image generations. However, most of the existing methods only generate face images from random noise, and cannot generate face images according to the specific attributes. In this paper, we focus on the problem of face synthesis from attributes, which aims at generating faces with specific characteristics corresponding to the given attributes. To this end, we propose a novel attributes aware face image generator method with generative adversarial networks called AFGAN. Specifically, we firstly propose a two-path embedding layer and self-attention mechanism to convert binary attribute vector to rich attribute features. Then three stacked generators generate $64 \times 64$, $128 \times 128$ and $256 \times 256$ resolution face images respectively by taking the attribute features as input. In addition, an image-attribute matching loss is proposed to enhance the correlation between the generated images and input attributes. Extensive experiments on CelebA demonstrate the superiority of our AFGAN in terms of both qualitative and quantitative evaluations.

cs.CV↗

arXiv2020

Delayed Feedback Modeling for the Entire Space Conversion Rate Prediction

Estimating post-click conversion rate (CVR) accurately is crucial in E-commerce. However, CVR prediction usually suffers from three major challenges in practice: i) data sparsity: compared with impressions, conversion samples are often extremely scarce; ii) sample selection bias: conventional CVR models are trained with clicked impressions while making inference on the entire space of all impressions; iii) delayed feedback: many conversions can only be observed after a relatively long and random delay since clicks happened, resulting in many false negative labels during training. Previous studies mainly focus on one or two issues while ignoring the others. In this paper, we propose a novel neural network framework ESDF to tackle the above three challenges simultaneously. Unlike existing methods, ESDF models the CVR prediction from a perspective of entire space, and combines the advantage of user sequential behavior pattern and the time delay factor. Specifically, ESDF utilizes sequential behavior of user actions on the entire space with all impressions to alleviate the sample selection bias problem. By sharing the embedding parameters between CTR and CVR networks, data sparsity problem is greatly relieved. Different from conventional delayed feedback methods, ESDF does not make any special assumption about the delay distribution. We discretize the delay time by day slot and model the probability based on survival analysis with deep neural network, which is more practical and suitable for industrial situations. Extensive experiments are conducted to evaluate the effectiveness of our method. To the best of our knowledge, ESDF is the first attempt to unitedly solve the above three challenges in CVR prediction area.

cs.LG↗

arXiv2020

Passport-aware Normalization for Deep Model Protection

Despite tremendous success in many application scenarios, deep learning faces serious intellectual property (IP) infringement threats. Considering the cost of designing and training a good model, infringements will significantly infringe the interests of the original model owner. Recently, many impressive works have emerged for deep model IP protection. However, they either are vulnerable to ambiguity attacks, or require changes in the target network structure by replacing its original normalization layers and hence cause significant performance drops. To this end, we propose a new passport-aware normalization formulation, which is generally applicable to most existing normalization layers and only needs to add another passport-aware branch for IP protection. This new branch is jointly trained with the target model but discarded in the inference stage. Therefore it causes no structure change in the target model. Only when the model IP is suspected to be stolen by someone, the private passport-aware branch is added back for ownership verification. Through extensive experiments, we verify its effectiveness in both image and 3D point recognition models. It is demonstrated to be robust not only to common attack techniques like fine-tuning and model compression, but also to ambiguity attacks. By further combining it with trigger-set based methods, both black-box and white-box verification can be achieved for enhanced security of deep learning models deployed in real systems. Code can be found at https://github.com/ZJZAC/Passport-aware-Normalization.

cs.CV↗

arXiv2020

FrUITeR: A Framework for Evaluating UI Test Reuse

UI testing is tedious and time-consuming due to the manual effort required. Recent research has explored opportunities for reusing existing UI tests from an app to automatically generate new tests for other apps. However, the evaluation of such techniques currently remains manual, unscalable, and unreproducible, which can waste effort and impede progress in this emerging area. We introduce FrUITeR, a framework that automatically evaluates UI test reuse in a reproducible way. We apply FrUITeR to existing test-reuse techniques on a uniform benchmark we established, resulting in 11,917 test reuse cases from 20 apps. We report several key findings aimed at improving UI test reuse that are missed by existing work.

cs.SE↗

arXiv2020

Trapping and acceleration of spin-polarized positrons from $γ$ photon splitting in wakefields

Energetic spin-polarized positrons are extremely demanded for forefront researches, such as $e^- e^+$ collider physics, but making compact positron sources is still very challenging. Here we put forward an efficient scheme of trapping and acceleration of polarized positrons in plasma wakefields. Seed electrons colliding with a bichromatic laser create polarized $γ$ photons which then split into $e^- e^+$ pairs via nonlinear Breit-Wheeler process with an average (partial) positron polarization above 30\% (70\%). Over 70\% positrons are trapped and accelerated in recovered wakefields driven by a hollow electron beam, obtaining an energy gain of 3.5 GeV/cm with slight depolarization. This scheme provides a potential for constructing compact and economical positron sources for future applications.

physics.plasm-ph↗

arXiv2020

The study of calibration for the hybrid pixel detector with single photon counting in HEPS-BPIX

The calibration process for the hybrid array pixel detector designed for High Energy Photon Source in China, we called HEPS-BPIX, is presented in this paper. Based on the threshold scanning, the relationship between energy and threshold is quantified for the threshold calibration. For the threshold trimming, the precise algorithm basing on LDAC characteristic and fast algorithm basing on LDAC scanning are proposed in this paper to study the performance of the threshold DACs which will be applied to the pixel. The threshold dispersion has been reduced from 46.28 mV without algorithm to 6.78 mV with the precise algorithm, whereas it is 7.61 mV with fast algorithm. For the temperature from 5 to 60 , the threshold dispersion of precise algorithm varies in the range of about 5.69 mV, whereas it is about 33.21 mV with the fast algorithm which can be re-corrected to 1.49 mV. The measurement results show that the fast algorithm could get the applicable threshold dispersion for a silicon pixel module and take a shorter time, while the precise algorithm could get better threshold dispersion, but time consuming. The temperature dependence of the silicon pixel module noise is also studied to assess the detector working status. The minimum detection energy can be reduced about 0.83 keV at a 20 lower temperature.

physics.ins-det↗

arXiv2020

Adsorption-controlled growth of MnTe(Bi2Te3)n by molecular beam epitaxy exhibiting stoichiometry-controlled magnetism

We report the growth of the intrinsic magnetic topological system MnTe(Bi2Te3)n by molecular beam epitaxy. By mapping the temperature and the Bi:Mn flux ratio, it is shown that there is a narrow growth window for the n=1 phase MnBi2Te4 with 2.0 4.5) it is found that the n = 2 MnBi4Te7 phase is stabilized. Transport measurements indicate that the MnBi2Te4 and MnBi4Te7 undergo magnetic transitions around 25 K, and 10 K, respectively, consistent with antiferromagnetic phases found in the bulk. Further, for Mn-rich conditions (Bi:Mn<2), ferromagnetism emerges that exhibits a clear hysteretic state in the Hall effect, which likely indicates Mn-doped MnBi2Te4. Understanding how to grow ternary chalcogenide phases is the key to synthesizing new materials and to interface magnetism and topology, which together are routes to realize and control exotic quantum phenomena.

cond-mat.mtrl-sci↗

arXiv2020

X-ray tomography investigation of cyclically sheared granular materials

We perform combined X-ray tomography and shear force measurements on a cyclically sheared granular system with highly transient behaviors, and obtain the evolution of microscopic structures and the macroscopic shear force during the shear cycle. We explain the macroscopic behaviors of the system based on microscopic processes, including the particle level structural rearrangement and frictional contact variation. Specifically, we show how contact friction can induce large structural fluctuations and cause significant shear dilatancy effect for granular materials, and we also construct an empirical constitutive relationship for the macroscopic shear force.

cond-mat.soft↗

arXiv2020

Learning to Dispatch for Job Shop Scheduling via Deep Reinforcement Learning

Priority dispatching rule (PDR) is widely used for solving real-world Job-shop scheduling problem (JSSP). However, the design of effective PDRs is a tedious task, requiring a myriad of specialized knowledge and often delivering limited performance. In this paper, we propose to automatically learn PDRs via an end-to-end deep reinforcement learning agent. We exploit the disjunctive graph representation of JSSP, and propose a Graph Neural Network based scheme to embed the states encountered during solving. The resulting policy network is size-agnostic, effectively enabling generalization on large-scale instances. Experiments show that the agent can learn high-quality PDRs from scratch with elementary raw features, and demonstrates strong performance against the best existing PDRs. The learned policies also perform well on much larger instances that are unseen in training.

cs.LG↗

arXiv2020

Low-frequency whistler waves excited by relativistic laser pulses

It is shown by multi-dimensional particle-in-cell simulations that intense secondary whistler waves with special vortex-like field topology can be excited by a relativistic laser pulse in the highly magnetized, near-critical density plasma. Such whistler waves with lower frequencies obliquely propagate on both sides of the laser propagation axis. The energy conversion rate from laser to whistler waves can exceed 15%. Their dispersion relations and field polarization properties can be well explained by the linear cold-plasma model. The present work presents a new excitation mechanism of whistler modes extending to the relativistic regime and could also be applied in magnetically assisted fast ignition.

physics.plasm-ph↗

arXiv2020

Frictional Granular Matter: Protocol Dependence of Mechanical Properties

Theoretical treatments of frictional granular matter often assume that it is legitimate to invoke classical elastic theory to describe its coarse-grained mechanical properties. Here we show, based on experiments and numerical simulations, that this is generically not the case since stress auto-correlation functions decay more slowly than the elastic Green's function. It was shown theoretically that standard elastic decay demands pressure and torque density fluctuations to be normal, with possibly one of them being hyperuniform. Generic compressed frictional assemblies exhibit however abnormal pressure fluctuations, failing to conform with the central limit theorem. The physics of this failure is linked to correlations built in the material during compression from a dilute configuration prior to jamming. By changing the protocol of compression one can observe different pressure fluctuations and stress auto-correlations decay at large scales.

cond-mat.soft↗

arXiv2020

Worm quantum Monte-Carlo study of phase diagram of extended Jaynes-Cummings-Hubbard model

Herein, we study the extended Jaynes-Cummings-Hubbard model mainly by the large-scale worm quantum Monte-Carlo method to check whether or not a light supersolid phase exists in various geometries, such as the one-dimensional chain, square lattices and triangular lattices. To achieve our purpose, the ground state phase diagrams are investigated. For the one-dimensional chain and square lattices, a first-order transition occurs between the superfluid phase and the solid phase and therefore there is no stable supersolid phase existing in these geometries. Interestingly, soliton/beats of the local densities arise if the chemical potential is adjusted in the finite-size chain. However, this soliton-superfluid coexistence can not be considered as a supersolid in the thermodynamic limit. Searching for a light supersolid, we also studied the Jaynes-Cummings-Hubbard model on triangular lattices, and the phase diagrams are obtained. Through measurement of the structural factor, momentum distribution and superfluid stiffness for various system sizes, a supersolid phase exists stably in the triangular lattices geometry and the regime of the supersolid phase is smaller than that of the mean field results. The light supersolid in the Jaynes-Cummings-Hubbard model is attractive because it has superreliance, which is absent in the pure Bose-Hubbard model. We believe the results in this paper could help search for new novel phases in cold-atom experiments

cond-mat.stat-mech↗

arXiv2020

Emergent Elasticity in Amorphous Solids

The mechanical response of naturally abundant amorphous solids such as gels, jammed grains, and biological tissues are not described by the conventional paradigm of broken symmetry that defines crystalline elasticity. In contrast, the response of such athermal solids are governed by local conditions of mechanical equilibrium, i.e., force and torque balance of its constituents. Here we show that these constraints have the mathematical structure of a generalized electromagnetism, where the electrostatic limit successfully captures the anisotropic elasticity of amorphous solids. The emergence of elasticity from local mechanical constraints offers a new paradigm for systems with no broken symmetry, analogous to emergent gauge theories of quantum spin liquids. Specifically, our $U(1)$ rank-2 symmetric tensor gauge theory of elasticity translates to the electromagnetism of fractonic phases of matter with the stress mapped to electric displacement and forces to vector charges. We corroborate our theoretical results with numerical simulations of soft frictionless disks in both two and three dimensions, and experiments on frictional disks in two dimensions. We also present experimental evidence indicating that force chains in granular media are sub-dimensional excitations of amorphous elasticity similar to fractons.

cond-mat.dis-nn↗

arXiv2020

Activity Complexes and A Prominent Poleward Surge During Solar Cycle 24

Long-lasting activity complexes (ACs), characterised as a series of closely located, continuously emerging solar active regions (ARs), are considered generating prominent poleward surges from observations. The surges lead to significant variations of the polar field, which are important for the modulation of solar cycles. We aim to study a prominent poleward surge during solar cycle 24 on the southern hemisphere, and analyse its originating ACs and the effect on the polar field evolution. We automatically identify and characterize ARs based on synoptic magnetograms from the Solar Dynamic Observatory. We assimilate these ARs with realistic magnetic configuration into a surface flux transport model, and simulate the creation and migration of the surge. Our simulations well reproduce the characteristics of the surge and show that the prominent surge is mainly caused by the ARs belonging to two ACs during Carrington Rotations 2145-2159 (December 2013-January 2015). The surge has a strong influence on the polar field evolution of the southern hemisphere during the latter half of cycle 24. Without the about one-year-long flux emergence in the form of ACs, the polar field around the cycle minimum would have remained at a low level and even reversed to the polarity at cycle 23 minimum. Our study also shows that the long-lived unipolar regions due to the decay of the earlier emerging ARs cause an intrinsic difficulty of automatically identifying and precisely quantifying later emerging ARs in ACs.

astro-ph.SR↗

arXiv2020

Privacy-Preserving Machine Learning Training in Aggregation Scenarios

To develop Smart City, the growing popularity of Machine Learning (ML) that appreciates high-quality training datasets generated from diverse IoT devices raises natural questions about the privacy guarantees that can be provided in such settings. Privacy-preserving ML training in an aggregation scenario enables a model demander to securely train ML models with the sensitive IoT data gathered from personal IoT devices. Existing solutions are generally server-aided, cannot deal with the collusion threat between the servers or between the servers and data owners, and do not match the delicate environments of IoT. We propose a privacy-preserving ML training framework named Heda that consists of a library of building blocks based on partial homomorphic encryption (PHE) enabling constructing multiple privacy-preserving ML training protocols for the aggregation scenario without the assistance of untrusted servers and defending the security under collusion situations. Rigorous security analysis demonstrates the proposed protocols can protect the privacy of each participant in the honest-but-curious model and defend the security under most collusion situations. Extensive experiments validate the efficiency of Heda which achieves the privacy-preserving ML training without losing the model accuracy.

cs.CR↗

arXiv2020

Cascaded Semantic and Positional Self-Attention Network for Document Classification

Transformers have shown great success in learning representations for language modelling. However, an open challenge still remains on how to systematically aggregate semantic information (word embedding) with positional (or temporal) information (word orders). In this work, we propose a new architecture to aggregate the two sources of information using cascaded semantic and positional self-attention network (CSPAN) in the context of document classification. The CSPAN uses a semantic self-attention layer cascaded with Bi-LSTM to process the semantic and positional information in a sequential manner, and then adaptively combine them together through a residue connection. Compared with commonly used positional encoding schemes, CSPAN can exploit the interaction between semantics and word positions in a more interpretable and adaptive manner, and the classification performance can be notably improved while simultaneously preserving a compact model size and high convergence rate. We evaluate the CSPAN model on several benchmark data sets for document classification with careful ablation studies, and demonstrate the encouraging results compared with state of the art.

cs.CL↗

arXiv2020

Federated Mutual Learning

Federated learning (FL) enables collaboratively training deep learning models on decentralized data. However, there are three types of heterogeneities in FL setting bringing about distinctive challenges to the canonical federated learning algorithm (FedAvg). First, due to the Non-IIDness of data, the global shared model may perform worse than local models that solely trained on their private data; Second, the objective of center server and clients may be different, where center server seeks for a generalized model whereas client pursue a personalized model, and clients may run different tasks; Third, clients may need to design their customized model for various scenes and tasks; In this work, we present a novel federated learning paradigm, named Federated Mutual Leaning (FML), dealing with the three heterogeneities. FML allows clients training a generalized model collaboratively and a personalized model independently, and designing their private customized models. Thus, the Non-IIDness of data is no longer a bug but a feature that clients can be personally served better. The experiments show that FML can achieve better performance than alternatives in typical FL setting, and clients can be benefited from FML with different models and tasks.

cs.LG↗

arXiv2020

Low-Rank Reorganization via Proportional Hazards Non-negative Matrix Factorization Unveils Survival Associated Gene Clusters

One of the central goals in precision health is the understanding and interpretation of high-dimensional biological data to identify genes and markers associated with disease initiation, development, and outcomes. Though significant effort has been committed to harness gene expression data for multiple analyses while accounting for time-to-event modeling by including survival times, many traditional analyses have focused separately on non-negative matrix factorization (NMF) of the gene expression data matrix and survival regression with Cox proportional hazards model. In this work, Cox proportional hazards regression is integrated with NMF by imposing survival constraints. This is accomplished by jointly optimizing the Frobenius norm and partial log likelihood for events such as death or relapse. Simulation results on synthetic data demonstrated the superiority of the proposed method, when compared to other algorithms, in finding survival associated gene clusters. In addition, using human cancer gene expression data, the proposed technique can unravel critical clusters of cancer genes. The discovered gene clusters reflect rich biological implications and can help identify survival-related biomarkers. Towards the goal of precision health and cancer treatments, the proposed algorithm can help understand and interpret high-dimensional heterogeneous genomics data with accurate identification of survival-associated gene clusters.

q-bio.QM↗