arXiv subjects

Kun Wang

Publications and source records attributed to Kun Wang.

At least 397 records · Page 22Linked to original sources

arXiv2022

A mm-Wave Patch Antenna with Broad Bandwidth and a Wide Angular Range

A novel mm-wave microstrip-fed patch antenna with broad bandwidth and wide angular coverage suitable for integration in planar arrays is designed, analyzed and verified by measurements. The antenna provides a bandwidth of 13.1% between 34.1 GHz and 38.9 GHz, which is achieved by a slotted multiple resonances microstrip patch and a matching circuit in microstrip technology. The antenna is built on RO3003 substrate with top and ground layers, which is low cost compared to other techniques. For simple integration with microstrip and frontend circuits, the feeding happens in the top layer with a microstrip coupling gap feed. The wide half power beamwidth is achieved by suitably designed parasitic patches for the first resonant mode. The second resonant mode has a wide half power beamwidth by default. The half power beamwidth is between 100° and 125° within the matched bandwidth, which is a very good value for a microstrip patch antenna radiating over a ground plane. The measured input impedance and radiation characteristic show very good agreement with simulation results.

eess.SP↗

arXiv2022

Development of the modified quasichemical model in the distinguishable-pair approximation for multiple compositions of short-range ordering

A binary solution with short-range ordering (SRO) exhibits characteristic solution thermodynamics. The Modified Quasichemical Model in the Pair Approximation (MQMPA) can effectively capture the thermodynamic features of a binary solution with a onefold SRO. However, once the SRO occurs at multiple compositions in a binary solution, the MQMPA is in principle inconvenient to treat the solution thermodynamics. It usually requires a large number of model parameters to fit the thermodynamic data of such a solution. The present work proposes the Modified Quasichemical Model in the Distinguishable-Pair Approximation (MQMDPA), which is a further improvement of the MQMPA. The MQMDPA can more realistically describe the solution thermodynamics with manifold SROs using fewer model parameters. It benefits from grouping the ordered pairs, which were assumed to be indistinguishable within the MQMPA, into several distinguishable types. Each kind of ordered pair has unique coordination numbers and pair energy, which is responsible for describing one of the SROs at the desired composition and strength. Interestingly, the MQMDPA can be completely transformed into the MQMPA when all kinds of ordered pairs are assigned the same pair energy and coordination numbers. The distinguishable pairs have thus become indistinguishable. Several real liquids with at least two observed SROs were successfully treated by the MQMDPA to demonstrate its effectiveness and reliability.

cond-mat.mtrl-sci↗

arXiv2022

LDoS attack detection method based on traffic time-frequency characteristics

For the traditional denial-of-service attack detection methods have complex algorithms and high computational overhead, which are difficult to meet the demand of online detection; and the experimental environment is mostly a simulation platform, which is difficult to deploy in real network environment, we propose a real network environment-oriented LDoS attack detection method based on the time-frequency characteristics of traffic data. All the traffic data flowing through the Web server is obtained through the acquisition storage system, and the detection data set is constructed using pre-processing; the simple features of the flow fragments are used as input, and the deep neural network is used to learn the time-frequency domain features of normal traffic features and generate reconstructed sequences, and the LDoS attack is discriminated based on the differences between the reconstructed sequences and the input data in the time-frequency domain. The experimental results show that the proposed method can accurately detect the attack features in the flow fragments in a very short time and achieve high detection accuracy for complex and diverse LDoS attacks; since only the statistical features of the packets are used, there is no need to parse the packet data, which can be adapted to different network environments.

cs.CR↗

arXiv2022

Network Traffic Anomaly Detection Method Based on Multi scale Residual Feature

To address the problem that traditional network traffic anomaly detection algorithms do not suffi-ciently mine potential features in long time domain, an anomaly detection method based on mul-ti-scale residual features of network traffic is proposed. The original traffic is divided into subse-quences of different time spans using sliding windows, and each subsequence is decomposed and reconstructed into data sequences of different levels using wavelet transform technique; the stacked autoencoder (SAE) constructs similar feature space using normal network traffic, and gen-erates reconstructed error vector using the difference between reconstructed samples and input samples in the similar feature space; the multi-path residual group is used to learn reconstructed error The traffic classification is completed by a lightweight classifier. The experimental results show that the detection performance of the proposed method for anomalous network traffic is sig-nificantly improved compared with traditional methods; it confirms that the longer time span and more S transformation scales have positive effects on discovering potential diversity information in the original network traffic.

cs.NI↗

arXiv2022

Fiber-optic multimode interference sensing: comprehensive characterization and its potential for strain-insensitive temperature sensing

A strain-insensitive temperature sensor based on multimode interference using standard multimode fibers (MMFs) is proposed according to the comprehensive study of the characteristics of the MMFs. The temperature and strain dependences on the core diameter, numerical aperture (NA), and the length of the MMF section in the single-mode--multimode--single-mode (SMS) fiber structure are investigated experimentally. The results indicate that the larger core diameter of the MMF leads to higher temperature sensitivity but lower strain sensitivity (absolute values); the higher NA does not influence the temperature sensitivity but results in higher absolute value of strain sensitivity; the longer MMF section brings lower temperature sensitivity but does not have an impact on strain sensitivity. These findings also contribute to the theoretical analysis of the length dependence in the SMS fiber sensors. Besides, the results of the characterization study show that the strain sensitivity is relatively low, which brings a possibility to develop a strain-insensitive temperature sensor. The proposed sensor is used for temperature sensing while the strain is constantly applied from 0 to 1100 $με$ with steps of 100 $με$. The measured results are consistent with the comprehensive study. The mean temperature sensitivity is 6.14 pm/$^{\circ}$C with a standard deviation of 0.39 pm/$^{\circ}$C, which proves that the proposed temperature sensor exhibits good stability and is insensitive to strain. We expect that these results will provide a profound guideline to fiber sensors based on multimode interference.

physics.optics↗

arXiv2022

ERGO: Event Relational Graph Transformer for Document-level Event Causality Identification

Document-level Event Causality Identification (DECI) aims to identify causal relations between event pairs in a document. It poses a great challenge of across-sentence reasoning without clear causal indicators. In this paper, we propose a novel Event Relational Graph TransfOrmer (ERGO) framework for DECI, which improves existing state-of-the-art (SOTA) methods upon two aspects. First, we formulate DECI as a node classification problem by constructing an event relational graph, without the needs of prior knowledge or tools. Second, ERGO seamlessly integrates event-pair relation classification and global inference, which leverages a Relational Graph Transformer (RGT) to capture the potential causal chain. Besides, we introduce edge-building strategies and adaptive focal loss to deal with the massive false positives caused by common spurious correlation. Extensive experiments on two benchmark datasets show that ERGO significantly outperforms previous SOTA methods (13.1% F1 gains on average). We have conducted extensive quantitative analysis and case studies to provide insights for future research directions (Section 4.8).

cs.CL↗

arXiv2022

Prompt for Extraction? PAIE: Prompting Argument Interaction for Event Argument Extraction

In this paper, we propose an effective yet efficient model PAIE for both sentence-level and document-level Event Argument Extraction (EAE), which also generalizes well when there is a lack of training data. On the one hand, PAIE utilizes prompt tuning for extractive objectives to take the best advantages of Pre-trained Language Models (PLMs). It introduces two span selectors based on the prompt to select start/end tokens among input texts for each role. On the other hand, it captures argument interactions via multi-role prompts and conducts joint optimization with optimal span assignments via a bipartite matching loss. Also, with a flexible prompt design, PAIE can extract multiple arguments with the same role instead of conventional heuristic threshold tuning. We have conducted extensive experiments on three benchmarks, including both sentence- and document-level EAE. The results present promising improvements from PAIE (3.5\% and 2.3\% F1 gains in average on three benchmarks, for PAIE-base and PAIE-large respectively). Further analysis demonstrates the efficiency, generalization to few-shot settings, and effectiveness of different extractive prompt tuning strategies. Our code is available at https://github.com/mayubo2333/PAIE.

cs.CL↗

arXiv2022

Towards Robust 2D Convolution for Reliable Visual Recognition

2D convolution (Conv2d), which is responsible for extracting features from the input image, is one of the key modules of a convolutional neural network (CNN). However, Conv2d is vulnerable to image corruptions and adversarial samples. It is an important yet rarely investigated problem that whether we can design a more robust alternative of Conv2d for more reliable feature extraction. In this paper, inspired by the recently developed learnable sparse transform that learns to convert the CNN features into a compact and sparse latent space, we design a novel building block, denoted by RConv-MK, to strengthen the robustness of extracted convolutional features. Our method leverages a set of learnable kernels of different sizes to extract features at different frequencies and employs a normalized soft thresholding operator to adaptively remove noises and trivial features at different corruption levels. Extensive experiments on clean images, corrupted images as well as adversarial samples validate the effectiveness of the proposed robust module for reliable visual recognition. The source codes are enclosed in the submission.

cs.CV↗

arXiv2022

INTERN: A New Learning Paradigm Towards General Vision

Enormous waves of technological innovations over the past several years, marked by the advances in AI technologies, are profoundly reshaping the industry and the society. However, down the road, a key challenge awaits us, that is, our capability of meeting rapidly-growing scenario-specific demands is severely limited by the cost of acquiring a commensurate amount of training data. This difficult situation is in essence due to limitations of the mainstream learning paradigm: we need to train a new model for each new scenario, based on a large quantity of well-annotated data and commonly from scratch. In tackling this fundamental problem, we move beyond and develop a new learning paradigm named INTERN. By learning with supervisory signals from multiple sources in multiple stages, the model being trained will develop strong generalizability. We evaluate our model on 26 well-known datasets that cover four categories of tasks in computer vision. In most cases, our models, adapted with only 10% of the training data in the target domain, outperform the counterparts trained with the full set of data, often by a significant margin. This is an important step towards a promising prospect where such a model with general vision capability can dramatically reduce our reliance on data, thus expediting the adoption of AI technologies. Furthermore, revolving around our new paradigm, we also introduce a new data system, a new architecture, and a new benchmark, which, together, form a general vision ecosystem to support its future development in an open and inclusive manner. See project website at https://opengvlab.shlab.org.cn .

cs.CV↗

arXiv2022

The temporal limits of predicting fault failure

Machine learning models using seismic emissions can predict instantaneous fault characteristics such as displacement in laboratory experiments and slow slip in Earth. Here, we address whether the acoustic emission (AE) from laboratory experiments contains information about near-future frictional behavior. The approach uses a convolutional encoder-decoder containing a transformer layer. We use as input progressively larger AE input time windows and progressively larger output friction time windows. The attention map from the transformer is used to interpret which regions of the AE contain hidden information corresponding to future frictional behavior. We find that very near-term predictive information is indeed contained in the AE signal, but farther into the future the predictions are progressively worse. Notably, information for predicting near future frictional failure and recovery are found to be contained in the AE signal. This first effort predicting future fault frictional behavior with machine learning will guide efforts for applications in Earth.

physics.geo-ph↗

arXiv2022

The modified quasichemical model in the Distinguishable-Pair Approximation for multicomponent solutions

The Modified Quasichemcial Model in the Distinguishable-Pair Approximation (MQMDPA) for manifold short-range orders in liquids has been successfully extended to multicomponent solutions. The extension is conducted by means of the geometrical interpolation method. Three types of interpolation models, namely Kohler, Toop and Chou, are introduced to initially formulate the pair interaction energies in ternary solutions by employing those in their constituent binary solutions. The pair energies can be expanded in terms of the pair fractions (configuration-dependent) or in terms of the coordination-equivalent fractions (composition-dependent). These methods are subsequently extended for use in multicomponent solutions. A general formalism for the combined Kohler-Toop model is employed to permit complete freedom of choice to treat any ternary subsystems with a symmetric or asymmetric model. Meanwhile, a general Chou model is also used to treat all ternary subsystems without any human interference to select symmetric or asymmetric components but only dependent upon the similarity and difference in properties from each two binary solutions with a co-member. Advantages and shortcomings are critically discussed regarding the utilization of different interpolation models.

cond-mat.other↗

arXiv2022

Angular momentum and parity projected multidimensionally constrained relativistic Hartree-Bogoliubov model

The nuclear deformations are of fundamental importance in nuclear physics. Recently we developed a multi-dimensionally constrained relativistic Hartree-Bogoliubov (MDCRHB) model, in which all multipole deformations respecting the $V_4$ symmetry can be considered self-consistently. In this work we extend this model by incorporating the angular momentum projection (AMP) and parity projection (PP) to restore the rotational and parity symmetries broken in the mean-field level. This projected-MDCRHB (p-MDCRHB) model enables us to connect certain nuclear spectra to exotic intrinsic shapes such as triangle or tetrahedron. We present the details of the method and an exemplary calculation for $^{12}$C. We develop a triangular moment constraint to generate the triangular configurations consisting of three $α$ clusters arranged as an equilateral triangle. The resulting $^{12}$C spectra are consistent with that from a triangular rigid rotor for large separations between the $α$ clusters. We also calculate the $B(E2)$ and $B(E3)$ values for low-lying states and find good agreement with the experiments.

nucl-th↗

arXiv2022

Physical Properties of 29 sdB+dM Eclipsing Binaries in Zwicky Transient Facility

The development of large-scale time-domain surveys provides an opportunity to study the physical properties as well as the evolutionary scenario of B-type subdwarfs (sdB) and M-type dwarfs (dM). Here, we obtained 33 sdB+dM eclipsing binaries based on the Zwicky Transient Facility (ZTF) light curves and $Gaia$ early data release 3 (EDR3) parallaxes. By using the PHOEBE code for light curve analysis, we obtain probability distributions for parameters of 29 sdB+dM. $R_1$, $R_2$, and $i$ are well determined, and the average uncertainty of mass ratio $q$ is 0.08. Our parameters are in good agreement with previous works if a typical mass of sdB is assumed. Based on parameters of 29 sdB+dM, we find that both the mass ratio $q$ and the companion's radius $R_2$ decrease with the shortening of the orbital period. For the three sdB+dMs with orbital periods less than 0.075 days, their companions are all brown dwarfs. The masses and radii of the companions satisfy the mass--radius relation for low-mass stars and brown dwarfs. Companions with radii between $0.12R_\odot$ and $0.15R_\odot$ seem to be missing in the observations. As more short-period sdB+dM eclipsing binaries are discovered and classified in the future with ZTF and $Gaia$, we will have more information to constrain the evolutionary ending of sdB+dM.

astro-ph.SR↗

arXiv2022

Nonlinear Optimal Guidance for Fixed-Time Impact on a Stationary Target

This paper is concerned with devising the nonlinear optimal guidance for intercepting a stationary target with a fixed impact time. According to Pontryagin's Maximum Principle (PMP), some optimality conditions for the solutions of the nonlinear optimal interception problem are established, and the structure of the corresponding optimal control is presented. By employing the optimality conditions, we formulate a parameterized system so that its solution space is the same as that of the nonlinear optimal interception problem. As a consequence, a simple propagation of the parameterized system, without using any optimization method, is sufficient to generate enough sampled data for the mapping from current state and time-to-go to the optimal guidance command. By virtue of the universal approximation theorem, a feedforward neural network, trained by the generated data, is able to represent the mapping from current state and time-to-go to the optimal guidance command. Therefore, the trained network eventually can generate fixed-impact-time nonlinear optimal guidance within a constant time. Finally, the developed nonlinear optimal guidance is exemplified and studied through simulations, showing that the nonlinear optimal guidance law performs better than existing interception guidance laws.

math.OC↗

arXiv2021

A smuon in the NMSSM confronted with the muon g-2 and SUSY searches

Motivated by the recent supersymmetry (SUSY) search results which prefer most SUSY particles heavy, and muon g-2 anomaly which prefers colorless SUSY particles light, we explore the status of a light smuon (the SUSY partner of left-handed muon lepton) in the Next-to-Minimal Supersymmetric Standard Model (NMSSM). Assuming colored SUSY particles be heavy, and considering numerous experimental constraints including muon g-2, SUSY searches, and dark matter, we scan the parameter space in the NMSSM with Z3-symmetry, checking the status of colorless SUSY particles and their possible mass order, and paying special attention to the smuon. After calculation and discussions, we draw the following conclusions: (i) The dominated SUSY contributions to muon g-2 come from the chargino-sneutrino loops, but the muon g-2 anomaly can also constrain the mass of smuon seriously because of the mass-degenerate relation between smuon and muon sneutrino. To interpret the muon g-2 anomaly at $1σ$ ($2σ$) level, the smuon need to be lighter than 1 TeV (1.8 TeV). (ii) When $\tildeχ^0_1$ is wino- or higgsino-dominated, the smuon can escape from the direct searches with low mass, e.g., 300 GeV. (iii) When smuon and $\tildeχ^0_1$ are mass-degenerate, the smuon can be as light as 200 GeV, while the $\tildeχ^0_1$ is usually bino-dominated, or some singlino-dominated, and its relic density can most likely reach the observed value, and the dominated annihilating mechanism is slepton annihilations. In addition, we also list several benchmark points for further studies.

hep-ph↗

arXiv2021

Physical Implementability of Linear Maps and Its Application in Error Mitigation

Completely positive and trace-preserving maps characterize physically implementable quantum operations. On the other hand, general linear maps, such as positive but not completely positive maps, which can not be physically implemented, are fundamental ingredients in quantum information, both in theoretical and practical perspectives. This raises the question of how well one can simulate or approximate the action of a general linear map by physically implementable operations. In this work, we introduce a systematic framework to resolve this task using the quasiprobability decomposition technique. We decompose a target linear map into a linear combination of physically implementable operations and introduce the physical implementability measure as the least amount of negative portion that the quasiprobability must pertain, which directly quantifies the cost of simulating a given map using physically implementable quantum operations. We show this measure is efficiently computable by semidefinite programs and prove several properties of this measure, such as faithfulness, additivity, and unitary invariance. We derive lower and upper bounds in terms of the Choi operator's trace norm and obtain analytic expressions for several linear maps of practical interests. Furthermore, we endow this measure with an operational meaning within the quantum error mitigation scenario: it establishes the lower bound of the sampling cost achievable via the quasiprobability decomposition technique. In particular, for parallel quantum noises, we show that global error mitigation has no advantage over local error mitigation.

quant-ph↗

arXiv2021

Mitigating Quantum Errors via Truncated Neumann Series

Quantum gates and measurements on quantum hardware are inevitably subject to hardware imperfections that lead to quantum errors. Mitigating such unavoidable errors is crucial to explore the power of quantum hardware better. In this paper, we propose a unified framework that can mitigate quantum gate and measurement errors in computing quantum expectation values utilizing the truncated Neumann series. The essential idea is to cancel the effect of quantum error by approximating its inverse via linearly combining quantum errors of different orders produced by sequential applications of the quantum devices with carefully chosen coefficients. Remarkably, the estimation error decays exponentially in the truncated order, and the incurred error mitigation overhead is independent of the system size, as long as the noise resistance of the quantum device is moderate. We numerically test this framework for different quantum errors and find that the computation accuracy is substantially improved. Our framework possesses several vital advantages: it mitigates quantum gate and measurement errors in a unified manner, it neither assumes any error structure nor requires the tomography procedure to completely characterize the quantum errors, and most importantly, it is scalable. These advantages empower our quantum error mitigation framework to be efficient and practical and extend the ability of near-term quantum devices to deliver quantum applications.

quant-ph↗

arXiv2021

A synchrotron-based kilowatt-level radiation source for EUV lithograph

A compact damping ring with limited circumference of about 160 m is proposed for producing kilowatt-level coherent EUV radiation. The electron bunch in the ring is modulated by a 257nm wavelength laser with the help of the angular dispersion induced micro-bunching method [C. Feng and Z. Zhao, Sci. Rep. 7, 4724 (2017)]. Coherent radiation at 13.5 nm with an average power of about 2.5 kW can be achieved with the state-of-the-art accelerator and laser technologies.

physics.acc-ph↗