arXiv subjects

Jun Liu

Publications and source records attributed to Jun Liu.

At least 739 records · Page 41Linked to original sources

arXiv2019

Multichannel signal detection in interference and noise when signal mismatch happens

In this paper, we consider the problem of detecting a multichannel signal in interference and noise when signal mismatch happens. We first propose two selective detectors, since their strong selectivity is preferred in some situations. However, these two detectors would not be suitable candidates if a robust detector is needed. To overcome this shortcoming, we then devise a tunable detector, which is parametrized by a non-negative scaling factor, referred to as the tunable parameter. By adjusting the tunable parameter, the proposed detector can smoothly change its capability in rejecting or robustly detecting a mismatch signal. Moreover, one selective detector and the tunable detector with an appropriate tunable parameter can provide nearly the same detection performance as existing detectors in the absence of signal mismatch. We obtain analytical expressions for the probabilities of detection (PDs) and probabilities of false alarm (PFAs) of the three proposed detectors, which are verified by Monte Carlo simulations.

eess.SP↗

arXiv2019

BERT4Rec: Sequential Recommendation with Bidirectional Encoder Representations from Transformer

Modeling users' dynamic and evolving preferences from their historical behaviors is challenging and crucial for recommendation systems. Previous methods employ sequential neural networks (e.g., Recurrent Neural Network) to encode users' historical interactions from left to right into hidden representations for making recommendations. Although these methods achieve satisfactory results, they often assume a rigidly ordered sequence which is not always practical. We argue that such left-to-right unidirectional architectures restrict the power of the historical sequence representations. For this purpose, we introduce a Bidirectional Encoder Representations from Transformers for sequential Recommendation (BERT4Rec). However, jointly conditioning on both left and right context in deep bidirectional model would make the training become trivial since each item can indirectly "see the target item". To address this problem, we train the bidirectional model using the Cloze task, predicting the masked items in the sequence by jointly conditioning on their left and right context. Comparing with predicting the next item at each position in a sequence, the Cloze task can produce more samples to train a more powerful bidirectional model. Extensive experiments on four benchmark datasets show that our model outperforms various state-of-the-art sequential models consistently.

cs.IR↗

arXiv2019

SleepGuardian: An RF-based Healthcare System Guarding Your Sleep from Afar

The ever accelerating process of urbanization urges more and more population into the swelling cities. While city residents are enjoying an entertaining life supported by advanced informatics techniques like 5G and cloud computing, the same technologies have also gradually deprived their sleep, which is crucial for their wellness. Therefore, sleep monitoring has drawn significant attention from both research and industry communities. In this article, we first review the sleep monitoring issue and point out three essential properties of an ideal sleep healthcare system, i.e., realtime guarding, fine-grained logging, and cost-effectiveness. Based on the analysis, we present SleepGuardian, a Radio Frequence (RF) based sleep healthcare system leveraging signal processing, edge computing and machine learning.SleepGuardian offers an offline sleep logging service and an online abnormality warning service. The offline service provides a fine-grained sleep log like timing and regularity of bed time, onset of sleep and night time awakenings. The online service keeps guarding the subject for any abnormal behaviors during sleep like intensive body twitches and a sudden seizure attack. Once an abnormality happens,it will automatically warn the designated contacts like a nearby emergency room or a closeby relative.We prototype SleepGuardian with low-cost WiFi devices and evaluate it in real scenarios. Experimental results demonstrate that SleepGuardian is very effective.

eess.SP↗

arXiv2019

A Regularized Convolutional Neural Network for Semantic Image Segmentation

Convolutional neural networks (CNNs) show outstanding performance in many image processing problems, such as image recognition, object detection and image segmentation. Semantic segmentation is a very challenging task that requires recognizing, understanding what's in the image in pixel level. Though the state of the art has been greatly improved by CNNs, there is no explicit connections between prediction of neighbouring pixels. That is, spatial regularity of the segmented objects is still a problem for CNNs. In this paper, we propose a method to add spatial regularization to the segmented objects. In our method, the spatial regularization such as total variation (TV) can be easily integrated into CNN network. It can help CNN find a better local optimum and make the segmentation results more robust to noise. We apply our proposed method to Unet and Segnet, which are well established CNNs for image segmentation, and test them on WBC, CamVid and SUN-RGBD datasets, respectively. The results show that the regularized networks not only could provide better segmentation results with regularization effect than the original ones but also have certain robustness to noise.

cs.CV↗

arXiv2019

NTU RGB+D 120: A Large-Scale Benchmark for 3D Human Activity Understanding

Research on depth-based human activity analysis achieved outstanding performance and demonstrated the effectiveness of 3D representation for action recognition. The existing depth-based and RGB+D-based action recognition benchmarks have a number of limitations, including the lack of large-scale training samples, realistic number of distinct class categories, diversity in camera views, varied environmental conditions, and variety of human subjects. In this work, we introduce a large-scale dataset for RGB+D human action recognition, which is collected from 106 distinct subjects and contains more than 114 thousand video samples and 8 million frames. This dataset contains 120 different action classes including daily, mutual, and health-related activities. We evaluate the performance of a series of existing 3D activity analysis methods on this dataset, and show the advantage of applying deep learning methods for 3D-based human action recognition. Furthermore, we investigate a novel one-shot 3D activity recognition problem on our dataset, and a simple yet effective Action-Part Semantic Relevance-aware (APSR) framework is proposed for this task, which yields promising results for recognition of the novel action classes. We believe the introduction of this large-scale dataset will enable the community to apply, adapt, and develop various data-hungry learning techniques for depth-based and RGB+D-based human activity understanding. [The dataset is available at: http://rose1.ntu.edu.sg/Datasets/actionRecognition.asp]

cs.CV↗

arXiv2019

Feature Boosting Network For 3D Pose Estimation

In this paper, a feature boosting network is proposed for estimating 3D hand pose and 3D body pose from a single RGB image. In this method, the features learned by the convolutional layers are boosted with a new long short-term dependence-aware (LSTD) module, which enables the intermediate convolutional feature maps to perceive the graphical long short-term dependency among different hand (or body) parts using the designed Graphical ConvLSTM. Learning a set of features that are reliable and discriminatively representative of the pose of a hand (or body) part is difficult due to the ambiguities, texture and illumination variation, and self-occlusion in the real application of 3D pose estimation. To improve the reliability of the features for representing each body part and enhance the LSTD module, we further introduce a context consistency gate (CCG) in this paper, with which the convolutional feature maps are modulated according to their consistency with the context representations. We evaluate the proposed method on challenging benchmark datasets for 3D hand pose estimation and 3D full body pose estimation. Experimental results show the effectiveness of our method that achieves state-of-the-art performance on both of the tasks.

cs.CV↗

arXiv2019

Fuel Economy Gaps Within & Across Garages: A Bivariate Random Parameters Seemingly Unrelated Regression Approach

The key objective of this study is to investigate the interrelationship between fuel economy gaps and to quantify the differential effects of several factors on fuel economy gaps of vehicles operated by the same garage. By using a unique fuel economy database (fueleconomy.gov), users self-reported fuel economy estimates and government fuel economy ratings are analyzed for more than 7000 garages across the U.S. The empirical analysis, nonetheless, is complicated owing to the presence of important methodological concerns including potential interrelationship between vehicles within the same garage and unobserved heterogeneity. To address these concerns, bivariate seemingly unrelated fixed and random parameter models are presented. With government test cycle ratings tending to over-estimate the actual on-road fuel economy, a significant variation is observed in the fuel economy gaps for the two vehicles across garages. A wide variety of factors such as driving style, fuel economy calculation method, and several vehicle specific characteristics are considered. Drivers who drive for maximum gas mileage or drives with the traffic flow have greater on-road fuel economy relative to the government official ratings. Contrarily, volatile drivers have smaller on-road fuel economy relative to the official ratings. Compared to the previous findings, our analysis suggests that the relationship between fuel type and fuel economy gaps is complex and not unidirectional. Regarding several vehicle and manufacturer related variables, the effects do not just significantly vary in magnitude but also in the direction, underscoring the importance of accounting for within-garage correlation and unobserved heterogeneity for making reliable inferences.

stat.AP↗

arXiv2019

Multi-dimensional entanglement transport through single-mode fibre

The global quantum network requires the distribution of entangled states over long distances, with significant advances already demonstrated using entangled polarisation states, reaching approximately 1200 km in free space and 100 km in optical fibre. Packing more information into each photon requires Hilbert spaces with higher dimensionality, for example, that of spatial modes of light. However spatial mode entanglement transport requires custom multimode fibre and is limited by decoherence induced mode coupling. Here we transport multi-dimensional entangled states down conventional single-mode fibre (SMF). We achieve this by entangling the spin-orbit degrees of freedom of a bi-photon pair, passing the polarisation (spin) photon down the SMF while accessing multi-dimensional orbital angular momentum (orbital) subspaces with the other. We show high fidelity hybrid entanglement preservation down 250 m of SMF across multiple 2x2 dimensions, demonstrating quantum key distribution protocols, quantum state tomographies and quantum erasers. This work offers an alternative approach to spatial mode entanglement transport that facilitates deployment in legacy networks across conventional fibre.

quant-ph↗

arXiv2019

Skeleton-Based Online Action Prediction Using Scale Selection Network

Action prediction is to recognize the class label of an ongoing activity when only a part of it is observed. In this paper, we focus on online action prediction in streaming 3D skeleton sequences. A dilated convolutional network is introduced to model the motion dynamics in temporal dimension via a sliding window over the temporal axis. Since there are significant temporal scale variations in the observed part of the ongoing action at different time steps, a novel window scale selection method is proposed to make our network focus on the performed part of the ongoing action and try to suppress the possible incoming interference from the previous actions at each step. An activation sharing scheme is also proposed to handle the overlapping computations among the adjacent time steps, which enables our framework to run more efficiently. Moreover, to enhance the performance of our framework for action prediction with the skeletal input data, a hierarchy of dilated tree convolutions are also designed to learn the multi-level structured semantic representations over the skeleton joints at each frame. Our proposed approach is evaluated on four challenging datasets. The extensive experiments demonstrate the effectiveness of our method for skeleton-based online action prediction.

cs.CV↗

arXiv2019

Robust Decidability of Sampled-Data Control of Nonlinear Systems with Temporal Logic Specifications

This paper explores the theoretical limits of using discrete abstractions for nonlinear control synthesis. More specifically, we consider the problem of deciding continuous-time control with temporal logic specifications. We prove that sampled-data control of nonlinear systems with temporal logic specifications is robustly decidable in the sense that, given a continuous-time nonlinear control system and a temporal logic formula, one can algorithmically decide whether there exists a robust sampled-data control strategy to realize this specification when the right-hand side of the system is slightly perturbed by a small disturbance. If the answer is positive, one can then construct a (potentially less) robust sampled-data control strategy that realizes the same specification. The result is proved by constructing a robustly complete abstraction of the original continuous-time control system using sufficiently small discretization parameters. We illustrate the result with a nonlinear control example.

eess.SY↗

arXiv2019

H2B: Heartbeat-based Secret Key Generation Using Piezo Vibration Sensors

We present Heartbeats-2-Bits (H2B), which is a system for securely pairing wearable devices by generating a shared secret key from the skin vibrations caused by heartbeat. This work is motivated by potential power saving opportunity arising from the fact that heartbeat intervals can be detected energy-efficiently using inexpensive and power-efficient piezo sensors, which obviates the need to employ complex heartbeat monitors such as Electrocardiogram or Photoplethysmogram. Indeed, our experiments show that piezo sensors can measure heartbeat intervals on many different body locations including chest, wrist, waist, neck and ankle. Unfortunately, we also discover that the heartbeat interval signal captured by piezo vibration sensors has low Signal-to-Noise Ratio (SNR) because they are not designed as precision heartbeat monitors, which becomes the key challenge for H2B. To overcome this problem, we first apply a quantile function-based quantization method to fully extract the useful entropy from the noisy piezo measurements. We then propose a novel Compressive Sensing-based reconciliation method to correct the high bit mismatch rates between the two independently generated keys caused by low SNR. We prototype H2B using off-the-shelf piezo sensors and evaluate its performance on a dataset collected from different body positions of 23 participants. Our results show that H2B has an overwhelming pairing success rate of 95.6%. We also analyze and demonstrate H2B's robustness against three types of attacks. Finally, our power measurements show that H2B is very power-efficient.

cs.CR↗

arXiv2019

Emergence of topologically protected states in MoTe$_{2}$ Weyl semimetal with layer stacking order

Electronic tunability in crystals with weakly-bound layers can be achieved through layer stacking order. One such example is MoTe$_2$, where the low-temperature orthorhombic T$_d$ phase is topological and host to Weyl quasiparticles. The transition mechanism to the non-trivial topology is elucidated by single crystal neutron diffraction. Upon cooling from the monoclinic 1T$\prime$ to the T$_d$ phase, diffuse scattering accompanies the transition, arising from random, in-plane layer displacements, and dissipates upon entering the T$_d$ phase. Diffuse scattering is observed only in the H0L plane due to irreversible layer shifts along the c-axis that break the centrosymmetry of the monoclinic lattice.

cond-mat.mtrl-sci↗

arXiv2019

Temporal contrast reduction techniques for high dynamic-range single-shot temporal contrast measurement

A single-shot characterization of the temporal contrast of a petawatt laser pulse with a high dynamic-range, is important not only for improving conditions of the petawatt laser facility itself, but also for various high-intensity laser physics experiments, which is still a difficult problem. In this study, a new idea for improving the dynamic-range of a single-shot temporal contrast measurement using novel temporal contrast reduction techniques is proposed. The proof-of-principle experiments applying single stage of pulse stretching, anti-saturated absorption, or optical Kerr effect successfully reduce the temporal contrast by approximately one order of magnitude. Combining with the SRSI-ETE method, its dynamic-range characterization capability is improved by approximately one order of magnitude to approximately 109. It is expected that a higher dynamic-range temporal contrast can be characterized using cascaded temporal contrast reduction processes.

physics.optics↗

arXiv2019

Highly Efficient Follicular Segmentation in Thyroid Cytopathological Whole Slide Image

In this paper, we propose a novel method for highly efficient follicular segmentation of thyroid cytopathological WSIs. Firstly, we propose a hybrid segmentation architecture, which integrates a classifier into Deeplab V3 by adding a branch. A large amount of the WSI segmentation time is saved by skipping the irrelevant areas using the classification branch. Secondly, we merge the low scale fine features into the original atrous spatial pyramid pooling (ASPP) in Deeplab V3 to accurately represent the details in cytopathological images. Thirdly, our hybrid model is trained by a criterion-oriented adaptive loss function, which leads the model converging much faster. Experimental results on a collection of thyroid patches demonstrate that the proposed model reaches 80.9% on the segmentation accuracy. Besides, 93% time is reduced for the WSI segmentation by using our proposed method, and the WSI-level accuracy achieves 53.4%.

cs.CV↗

arXiv2019

Polarization sensitive laser intensity inside femtosecond filament in air

Polarization dependence on clamping intensity inside femtosecond filament was experimentally measured in air. By tuning the laser pulse ellipse from linear polarization to circular polarization, the measured clamping intensity inside laser filament is gradually increased up to 1.7 times. The experimental results are in good agreement with the simulation results by solving the extended nonlinear Schrodinger equation (NLSE). The polarization sensitive clamping intensity inside filaments offers an important factor towards fully understanding the polarization related phenomenon observed so far.

physics.optics↗

arXiv2019

Ruling out $\sim 100-300$ GeV thermal relic annihilating dark matter by radio observation of the Andromeda galaxy

In the past few years, some studies claimed that annihilating dark matter with mass $\sim 10-100$ GeV can explain the GeV gamma-ray excess in our Galaxy. However, recent analyses of the Fermi-LAT and radio observational data rule out the possibility of the thermal relic annihilating dark matter with mass $m \le 100$ GeV for some popular annihilation channels. By using the new observed radio data of the Andromeda galaxy, we rule out the existence of $\sim 100-300$ GeV thermal relic annihilating dark matter for ten annihilation channels. The lower limits of annihilating dark matter mass are improved to larger than 330 GeV for the most conservative case, which is a few times larger than the current best constraints. Moreover, these limits strongly disfavor the benchmark model of weakly interacting massive particle (WIMP) produced through the thermal freeze-out mechanism.

astro-ph.GA↗

arXiv2019

Generation of high-energy clean multicolored ultrashort pulses and their application in single-shot temporal contrast measurement

We demonstrate the generation of 100-μJ-level multicolored femtosecond pulses based on a single-stage cascaded four-wave mixing (CFWM) process in a thin glass plate. The generated high-energy CFWM signals can shift the central wavelength and have well-enhanced temporal contrast because of the third-order nonlinear process. They are innovatively used as clean sampling pulses of a cross-correlator for single-shot temporal contrast measurement. With a simple home-made setup, the proof-of-principle experimental results demonstrate the single-shot cross-correlator with dynamic range of 1010, temporal resolution of about 160 fs and temporal window of 50 ps. To the best of our knowledge, this is the first demonstration in which both the dynamic range and the temporal resolution of a single-shot temporal contrast measurement are comparable to those of a commercial delay-scanning cross-correlator.

physics.optics↗

arXiv2018

Deep Neural Network Aided Scenario Identification in Wireless Multi-path Fading Channels

This letter illustrates our preliminary works in deep nerual network (DNN) for wireless communication scenario identification in wireless multi-path fading channels. In this letter, six kinds of channel scenarios referring to COST 207 channel model have been performed. 100% identification accuracy has been observed given signal-to-noise (SNR) over 20dB whereas a 88.4% average accuracy has been obtained where SNR ranged from 0dB to 40dB. The proposed method has tested under fast time-varying conditions, which were similar with real world wireless multi-path fading channels, enabling it to work feasibly in practical scenario identification.

eess.SP↗