arXiv subjects

Chen Wang

Publications and source records attributed to Chen Wang.

At least 703 records · Page 39Linked to original sources

arXiv2020

SAG-VAE: End-to-end Joint Inference of Data Representations and Feature Relations

Variational Autoencoders (VAEs) are powerful in data representation inference, but it cannot learn relations between features with its vanilla form and common variations. The ability to capture relations within data can provide the much needed inductive bias necessary for building more robust Machine Learning algorithms with more interpretable results. In this paper, inspired by recent advances in relational learning using Graph Neural Networks, we propose the Self-Attention Graph Variational AutoEncoder (SAG-VAE) network which can simultaneously learn feature relations and data representations in an end-to-end manner. SAG-VAE is trained by jointly inferring the posterior distribution of two types of latent variables, which denote the data representation and a shared graph structure, respectively. Furthermore, we introduce a novel self-attention graph network that improves the generative capabilities of SAG-VAE by parameterizing the generative distribution allowing SAG-VAE to generate new data via graph convolution, while still trainable via backpropagation. A learnable relational graph representation enhances SAG-VAE's robustness to perturbation and noise, while also providing deeper intuition into model performance. Experiments based on graphs show that SAG-VAE is capable of approximately retrieving edges and links between nodes based entirely on feature observations. Finally, results on image data illustrate that SAG-VAE is fairly robust against perturbations in image reconstruction and sampling.

cs.LG↗

arXiv2020

MSA-MIL: A deep residual multiple instance learning model based on multi-scale annotation for classification and visualization of glomerular spikes

Membranous nephropathy (MN) is a frequent type of adult nephrotic syndrome, which has a high clinical incidence and can cause various complications. In the biopsy microscope slide of membranous nephropathy, spikelike projections on the glomerular basement membrane is a prominent feature of the MN. However, due to the whole biopsy slide contains large number of glomeruli, and each glomerulus includes many spike lesions, the pathological feature of the spikes is not obvious. It thus is time-consuming for doctors to diagnose glomerulus one by one and is difficult for pathologists with less experience to diagnose. In this paper, we establish a visualized classification model based on the multi-scale annotation multi-instance learning (MSA-MIL) to achieve glomerular classification and spikes visualization. The MSA-MIL model mainly involves three parts. Firstly, U-Net is used to extract the region of the glomeruli to ensure that the features learned by the succeeding algorithm are focused inside the glomeruli itself. Secondly, we use MIL to train an instance-level classifier combined with MSA method to enhance the learning ability of the network by adding a location-level labeled reinforced dataset, thereby obtaining an example-level feature representation with rich semantics. Lastly, the predicted scores of each tile in the image are summarized to obtain glomerular classification and visualization of the classification results of the spikes via the usage of sliding window method. The experimental results confirm that the proposed MSA-MIL model can effectively and accurately classify normal glomeruli and spiked glomerulus and visualize the position of spikes in the glomerulus. Therefore, the proposed model can provide a good foundation for assisting the clinical doctors to diagnose the glomerular membranous nephropathy.

cs.CV↗

arXiv2020

Visual Memorability for Robotic Interestingness via Unsupervised Online Learning

In this paper, we explore the problem of interesting scene prediction for mobile robots. This area is currently underexplored but is crucial for many practical applications such as autonomous exploration and decision making. Inspired by industrial demands, we first propose a novel translation-invariant visual memory for recalling and identifying interesting scenes, then design a three-stage architecture of long-term, short-term, and online learning. This enables our system to learn human-like experience, environmental knowledge, and online adaption, respectively. Our approach achieves much higher accuracy than the state-of-the-art algorithms on challenging robotic interestingness datasets.

cs.CV↗

arXiv2020

Brownian thermal transistors and refrigerators in mesoscopic systems

Fluctuations are significant in mesoscopic systems and of particular importance in understanding quantum transport. Here, we show that fluctuations can be considered as a resource for the operations of open quantum systems as functional devices. We derive the statistics of the thermal transistor amplification factor and the cooling-by-heating refrigerator efficiency under the Gaussian fluctuation framework. Statistical properties of the stochastic thermal transistor and the cooling-by-heating efficiency are revealed in the linear-response regime. We clarify the unique role of inelastic processes on thermal transport in mesoscopic systems. We further show that elastic and inelastic processes lead to different bounds based on the linear transport coefficients by establishing a generic theoretical framework for mesoscopic heat transport, which treats electron and bosonic collective excitations in an equal-footing manner. The underlying physics are illustrated concretely using a double-quantum-dot three-terminal system, though the theory applies to more general systems.

cond-mat.mes-hall↗

arXiv2020

Filament Intersections and Cold Dense Cores in Orion A North

We studied the filament structures and dense cores in OMC-2,3 region in Orion A North molecular cloud using the high-resolution N2H+ (1-0) spectral cube observed with the Atacama Large Millimeter/Submillimeter Array (ALMA). The filament network over a total length of 2 pc is found to contain 170 intersections and 128 candidate dense cores. The dense cores are all displaced from the infrared point sources (possible young stars), and the major fraction of cores (103) are located around the intersections. Towards the intersections, there is also an increasing trend for the total column density Ntot as well as the the power-law index of the column-density Probability Distribution Function (N-PDF), suggesting that the intersections would in general have more significant gas assembly than the other part of the filament paths. The virial analysis shows that the dense cores mostly have virial mass ratio of alpha_vir=M_vir/M_gas<1.0, suggesting them to be bounded by the self gravity. In the mean time, only about 23 percent of the cores have critical mass ratio of alpha_crit=M_crit/M_gas<1.0, suggesting them to be unstable against core collapse. Combining these results, it shows that the major fraction of the cold starless and possible prestellar cores in OMC-2,3 are being assembled around the intersections, and currently in a gravitationally bound state. But more extensive core collapse and star formation may still require continuous core-mass growth or other perturbatio

astro-ph.GA↗

arXiv2020

Distances and statistics of local molecular clouds in the first Galactic quadrant

We present an analysis of local molecular clouds (-6 <VLSR< 30 km/s, i.e., <1.5 kpc) in the first Galactic quadrant (25.8° <l<49.7° and |b|<5°), a pilot region of the Milky Way Imaging Scroll Painting (MWISP) CO survey. Using the SCIMES algorithm to divide large molecular clouds into moderate-sized ones, we determined distances to 28 molecular clouds with the background-eliminated extinction-parallax (BEEP) method using the Gaia DR2 parallax measurements aided by AG and AV, and the distance ranges from 250 pc to about 1.5 kpc. These incomplete distance samples indicate a linear relationship between the distance and the radial velocity (VLSR) with a scatter of 0.16 kpc, and kinematic distances may be systematically larger for local molecular clouds. In order to investigate fundamental properties of molecular clouds, such as the total sample number, the linewidth, the brightness temperature, the physical area, and the mass, we decompose the spectral cube using the DBSCAN algorithm. Post selection criteria are imposed on DBSCAN clusters to remove the noise contamination, and we found that the separation of molecular cloud individuals is reliable based on a definition of independent consecutive structures in l-b-V space. The completeness of the local molecular cloud flux collected by the MWISP CO survey is about 80%. The physical area, A, shows a power-law distribution, dN/dA \propto A^{-2.20+/-0.18}, while the molecular cloud mass also follows a power-law distribution but slightly flatter, dN/dM \propto M^{-1.96+/-0.11}.

astro-ph.GA↗

arXiv2020

Time Series Data Cleaning with Regular and Irregular Time Intervals

Errors are prevalent in time series data, especially in the industrial field. Data with errors could not be stored in the database, which results in the loss of data assets. Handling the dirty data in time series is non-trivial, when given irregular time intervals. At present, to deal with these time series containing errors, besides keeping original erroneous data, discarding erroneous data and manually checking erroneous data, we can also use the cleaning algorithm widely used in the database to automatically clean the time series data. This survey provides a classification of time series data cleaning techniques and comprehensively reviews the state-of-the-art methods of each type. In particular, we have a special focus on the irregular time intervals. Besides we summarize data cleaning tools, systems and evaluation criteria from research and industry. Finally, we highlight possible directions time series data cleaning.

cs.DB↗

arXiv2020

A new kind of Hermite-Gaussian-like optical vortex generated by cross-phase

We propose a new kind of optical vortex called the Hermite-Gaussian-like optical vortex (HGOV) inspired by the crossphase (CP). Theoretically, we investigate how the CP is decoupled from the phase of a cylindrical lens. We also investigate the propagation characteristics of an HGOV, which has a Hermitian-Gaussian-like intensity distribution but still retains the orbital angular momentum. Furthermore, we derive the Fresnel diffraction integral of an HGOV and study the purity at infinity. Besides, we show a novel function of the self-measurement of the HGOV. Finally, we show that we can change the relative positions of singularities and the direction of an HGOV precisely, which facilitates applications in optical micromanipulation.

physics.optics↗

arXiv2020

On two conjectural supercongruences of Z.-W. Sun

In this paper, we mainly prove two conjectural supercongruences of Sun by using the following identity $$ \sum_{k=0}^n\binom{2k}{k}^2\binom{2n-2k}{n-k}^2=16^n\sum_{k=0}^n\frac{\binom{n+k}{k}\binom{n}{k}\binom{2k}{k}^2}{(-16)^k} $$ which arises from a ${}_4F_3$ hypergeometric transformation. For any prime $p>3$, we prove that \begin{gather*} \sum_{n=0}^{p-1}\frac{n+1}{8^n}\sum_{k=0}^n\binom{2k}{k}^2\binom{2n-2k}{n-k}^2\equiv(-1)^{(p-1)/2}p+5p^3E_{p-3}\pmod{p^4},\\ \sum_{n=0}^{p-1}\frac{2n+1}{(-16)^n}\sum_{k=0}^n\binom{2k}{k}^2\binom{2n-2k}{n-k}^2\equiv(-1)^{(p-1)/2}p+3p^3E_{p-3}\pmod{p^4}, \end{gather*} where $E_{p-3}$ is the $(p-3)$th Euler number.

math.NT↗

arXiv2020

Symbolic summation methods and hypergeometric supercongruences

In this paper, we establish the following two congruences: \begin{gather*} \sum_{k=0}^{(p+1)/2}(3k-1)\frac{\left(-\frac{1}{2}\right)_k^2\left(\frac{1}{2}\right)_k4^k}{k!^3}\equiv p-6p^3\left(\frac{-1}{p}\right)+2p^3\left(\frac{-1}{p}\right)E_{p-3}\pmod{p^4},\\ \sum_{k=0}^{p-1}(3k-1)\frac{\left(-\frac{1}{2}\right)_k^2\left(\frac{1}{2}\right)_k4^k}{k!^3}\equiv p-2p^3\pmod{p^4}, \end{gather*} where $p>3$ is a prime, $E_{p-3}$ is the $(p-3)$-th Euler number and $\left(-\right)$ is the Legendre symbol. The first congruence modulo $p^3$ was conjectured by Guo and Schlosser recently.

math.NT↗

arXiv2020

Properties of OB star-black hole systems derived from detailed binary evolution models

The recent gravitational wave measurements have demonstrated the existence of stellar mass black hole binaries. It is essential for our understanding of massive star evolution to identify the contribution of binary evolution to the formation of double black holes. A promising way to progress is investigating the progenitors of double black hole systems and comparing predictions with local massive star samples such as the population in 30 Doradus in the Large Magellanic Cloud (LMC). Methods. To this purpose, we analyse a large grid of detailed binary evolution models at LMC metallicity with initial primary masses between 10 and 40 Msun, and identify which model systems potentially evolve into a binary consisting of a black hole and a massive main sequence star. We then derive the observable properties of such systems, as well as peculiarities of the OB star component. We find that about 3% of the LMC late O and early B stars in binaries are expected to possess a black hole companion, when assuming stars with a final helium core mass above 6.6 M to form black holes. While the vast majority of them may be X-ray quiet, our models suggest that these may be identified in spectroscopic binaries, either by large amplitude radial velocity variations ( > 50 km s ) and simultaneous nitrogen surface enrichment, or through a moderate radial velocity ( > 10 km/s ) and simultaneously rapid rotation of the OB star. The predicted mass ratios are such that main sequence companions could be excluded in most cases. A comparison to the observed OB+WR binaries in the LMC, Be/X-ray binaries, and known massive BH binaries supports our conclusion. We expect spectroscopic observations to be able to test key assumptions in our models, with important implications for massive star evolution in general, and for the formation of double-black hole mergers in particular.

astro-ph.SR↗

arXiv2020

EchoLock: Towards Low Effort Mobile User Identification

User identification plays a pivotal role in how we interact with our mobile devices. Many existing authentication approaches require active input from the user or specialized sensing hardware, and studies on mobile device usage show significant interest in less inconvenient procedures. In this paper, we propose EchoLock, a low effort identification scheme that validates the user by sensing hand geometry via commodity microphones and speakers. These acoustic signals produce distinct structure-borne sound reflections when contacting the user's hand, which can be used to differentiate between different people based on how they hold their mobile devices. We process these reflections to derive unique acoustic features in both the time and frequency domain, which can effectively represent physiological and behavioral traits, such as hand contours, finger sizes, holding strength, and gesture. Furthermore, learning-based algorithms are developed to robustly identify the user under various environments and conditions. We conduct extensive experiments with 20 participants using different hardware setups in key use case scenarios and study various attack models to demonstrate the performance of our proposed system. Our results show that EchoLock is capable of verifying users with over 90% accuracy, without requiring any active input from the user.

cs.HC↗

arXiv2020

WearID: Wearable-Assisted Low-Effort Authentication to Voice Assistants using Cross-Domain Speech Similarity

Due to the open nature of voice input, voice assistant (VA) systems (e.g., Google Home and Amazon Alexa) are under a high risk of sensitive information leakage (e.g., personal schedules and shopping accounts). Though the existing VA systems may employ voice features to identify users, they are still vulnerable to various acoustic attacks (e.g., impersonation, replay and hidden command attacks). In this work, we focus on the security issues of the emerging VA systems and aim to protect the users' highly sensitive information from these attacks. Towards this end, we propose a system, WearID, which uses an off-the-shelf wearable device (e.g., a smartwatch or bracelet) as a secure token to verify the user's voice commands to the VA system. In particular, WearID exploits the readily available motion sensors from most wearables to describe the command sound in vibration domain and check the received command sound across two domains (i.e., wearable's motion sensor vs. VA device's microphone) to ensure the sound is from the legitimate user.

cs.HC↗

arXiv2020

WiEat: Fine-grained Device-free Eating Monitoring Leveraging Wi-Fi Signals

Eating is a fundamental activity in people's daily life. Studies have shown that many health-related problems such as obesity, diabetes and anemia are closely associated with people's unhealthy eating habits (e.g., skipping meals, eating irregularly and overeating). Traditional eating monitoring solutions relying on self-reports remain an onerous task, while the recent trend requiring users to wear expensive dedicated hardware is still invasive. To overcome these limitations, in this paper, we develop a device-free eating monitoring system using WiFi-enabled devices (e.g., smartphone or laptop). Our system aims to automatically monitor users' eating activities through identifying the fine-grained eating motions and detecting the chewing and swallowing. In particular, our system extracts the fine-grained Channel State Information (CSI) from WiFi signals to distinguish eating from non-eating activities and further recognizing users' detailed eating motions with different utensils (e.g., using a folk, knife, spoon or bare hands). Moreover, the system has the capability of identifying chewing and swallowing through detecting users' minute facial muscle movements based on the derived CSI spectrogram. Such fine-grained eating monitoring results are beneficial to the understanding of the user's eating behaviors and can be used to estimate food intake types and amounts. Extensive experiments with 20 users over 1600-minute eating show that the proposed system can recognize the user's eating motions with up to 95% accuracy and estimate the chewing and swallowing amount with 10% percentage error.

cs.HC↗

arXiv2020

Some $q$-congruences arising from certain identities

In this paper, by constructing some identities, we prove some $q$-analogues of some congruences. For example, for any odd integer $n>1$, we show that \begin{gather*} \sum_{k=0}^{n-1} \frac{(q^{-1};q^2)_k}{(q;q)_k} q^k \equiv (-1)^{(n+1)/2} q^{(n^2-1)/4} - (1+q)[n] \pmod{Φ_n(q)^2},\\ \sum_{k=0}^{n-1}\frac{(q^3;q^2)_k}{(q;q)_k} q^k \equiv (-1)^{(n+1)/2} q^{(n^2-9)/4} + \frac{1+q}{q^2}[n]\pmod{Φ_n(q)^2}, \end{gather*} where the $q$-Pochhanmmer symbol is defined by $(x;q)_0=1$ and $(x;q)_k = (1-x)(1-xq)\cdots(1-xq^{k-1})$ for $k\geq1$, the $q$-integer is defined by $[n]=1+q+\cdots+q^{n-1}$ and $Φ_n(q)$ is the $n$-th cyclotomic polynomial. The $q$-congruences above confirm some recent conjectures of Gu and Guo.

math.NT↗

arXiv2020

Methodology and Performance of the Two-Year Galactic Plane Scanning Survey of Insight-HXMT

The Galactic plane scanning survey is one of the main scientific objectives of the Hard X-ray Modulation Telescope (known as Insight-HXMT). During the two-year operation of Insight-HXMT, more than 1000 scanning observations have been performed and the whole Galactic plane ($\rm 0^{\circ}<l<360^{\circ}$, $\rm -10^{\circ}<b<10^{\circ}$) has been covered completely. We summarize the Galactic plane scanning survey of Insight-HXMT for two years, including the characteristics of the scanning data, the data analysis process and the preliminary results of the Low-Energy telescope, the Medium-Energy telescope and the High-Energy telescope. With the light curve PSF fitting method, the fluxes of the known sources in the scanned area as well as the flux errors are obtained for each scanning observation. From the relationships of SNRs and fluxes, the $5σ$ sensitivities of three telescopes of Insight-HXMT are estimated as $\rm \sim7.6\times10^{-11}~erg cm^{-2}~s^{-1}$ ($\rm 3 mCrab,~1-6 keV$), $\rm \sim4.0\times10^{-10}~erg~cm^{-2}~s^{-1}$ ($\rm 20~mCrab,~7-40 keV$) and $\rm \sim2.6\times10^{-10}~erg cm^{-2}~s^{-1}$ ($\rm 18 mCrab,~25-100 keV$) for an individual scanning observation of $2-3$ hours, respectively. Up to September 2019, more than 800 X-ray sources with various types are monitored by the three telescopes and their long-term light curves with three energy bands are obtained to make further scientific analyses.

astro-ph.IM↗

arXiv2020

Managing quantum heat transfer in nonequilibrium qubit-phonon hybrid system

We investigate quantum heat transfer and thermal management in the nonequilibrium qubit-phonon hybrid system by applying the quantum master equation embedded with phononic coherent state. We obtain the steady state heat flow by tuning the arbitrary qubit-phonon coupling strength, which particularly exhibits the power-law scaling behavior and the turnover behavior in the weak and strong coupling regimes, respectively. Moreover, we analyze the negative differential thermal conductance and thermal rectification, which becomes profound with weak qubit-phonon interaction and large temperature bias. These results would contribute to smart energy control and design of phononic hybrid quantum devices.

quant-ph↗

arXiv2020

Intensity Scan Context: Coding Intensity and Geometry Relations for Loop Closure Detection

Loop closure detection is an essential and challenging problem in simultaneous localization and mapping (SLAM). It is often tackled with light detection and ranging (LiDAR) sensor due to its view-point and illumination invariant properties. Existing works on 3D loop closure detection often leverage the matching of local or global geometrical-only descriptors, but without considering the intensity reading. In this paper we explore the intensity property from LiDAR scan and show that it can be effective for place recognition. Concretely, we propose a novel global descriptor, intensity scan context (ISC), that explores both geometry and intensity characteristics. To improve the efficiency for loop closure detection, an efficient two-stage hierarchical re-identification process is proposed, including a binary-operation based fast geometric relation retrieval and an intensity structure re-identification. Thorough experiments including both local experiment and public datasets test have been conducted to evaluate the performance of the proposed method. Our method achieves higher recall rate and recall precision than existing geometric-only methods.

cs.RO↗