arXiv subjects

Jun Li

Publications and source records attributed to Jun Li.

At least 577 records · Page 32Linked to original sources

arXiv2022

Faithful coherent states

We propose the notion of faithful coherent states based on the fidelity-based coherence witness. The criterion for detecting faithful coherent states can be restricted to a subclass of fidelity-based criterion under unitary transformations for single and bipartite systems. We can realize these unitary transformations by using quantum gates and circuits, and establish the connection of faithful coherence states and coherence distillation, maximum relative entropy of coherence.

quant-ph↗

arXiv2022

Measuring the primordial gravitational waves from cosmic microwave background and stochastic gravitational wave background observations

We constrain the primordial gravitational waves from cosmic microwave background (CMB) and stochastic gravitational wave background (SGWB) observations. SGWB provides the latest way to explore the early universe and the cosmological evolution which can be reflected by primordial gravitational waves. We not only combine LIGO observations with CMB to measure primordial gravitational waves, but also forecast the potential abilities of the LISA detector and PTA projects. In the $Λ$CDM+$r$+$n_t$ model, the standard six parameters change slightly from SGWB observations. While the constraints on tensor-to-scalar ratio and tensor spectral index are improved obviously from SGWB observations. FAST projects have a significant impact on tensor-to-scalar ratio and tensor spectral index, namely $r<0.028$ and $n_t=-0.41^{+0.64}_{-0.96}$ at $95\%$ confidence level.

astro-ph.CO↗

arXiv2022

Nested Collaborative Learning for Long-Tailed Visual Recognition

The networks trained on the long-tailed dataset vary remarkably, despite the same training settings, which shows the great uncertainty in long-tailed learning. To alleviate the uncertainty, we propose a Nested Collaborative Learning (NCL), which tackles the problem by collaboratively learning multiple experts together. NCL consists of two core components, namely Nested Individual Learning (NIL) and Nested Balanced Online Distillation (NBOD), which focus on the individual supervised learning for each single expert and the knowledge transferring among multiple experts, respectively. To learn representations more thoroughly, both NIL and NBOD are formulated in a nested way, in which the learning is conducted on not just all categories from a full perspective but some hard categories from a partial perspective. Regarding the learning in the partial perspective, we specifically select the negative categories with high predicted scores as the hard categories by using a proposed Hard Category Mining (HCM). In the NCL, the learning from two perspectives is nested, highly related and complementary, and helps the network to capture not only global and robust features but also meticulous distinguishing ability. Moreover, self-supervision is further utilized for feature enhancement. Extensive experiments manifest the superiority of our method with outperforming the state-of-the-art whether by using a single model or an ensemble.

cs.CV↗

arXiv2022

Almost global smooth solutions of the 3D quasilinear Klein-Gordon equations on the product space $\mathbb{R}^{2}\times \mathbb{T}$

In the paper, for the 3D quasilinear Klein-Gordon equation with the small initial data posed on the product space $\mathbb{R}^{2}\times \mathbb{T}$, we focus on the lower bound of the lifespan of the smooth solution. When the size of initial data is bounded by $\varepsilon_0>0$, by the space-time resonance method, it is shown that smooth solution exists up to the time $e^{c_{0}/\varepsilon_{0}^2}$ with $\varepsilon_0$ being sufficiently small and $c_0>0$ being some suitable constant.

math.AP↗

arXiv2022

On the blowup mechanism of smooth solutions to 1D quasilinear strictly hyperbolic systems with large initial data

For the first order 1D $n\times n$ quasilinear strictly hyperbolic system $\partial_tu+F(u)\partial_xu=0$ with $u(x, 0)=\varepsilon u_0(x)$, where $\varepsilon>0$ is small, $u_0(x)\not\equiv 0$ and $u_0(x)\in C_0^2(\mathbb R)$, when at least one eigenvalue of $F(u)$ is genuinely nonlinear, it is well-known that on the finite blowup time $T_{\varepsilon}$, the derivatives $\partial_{t,x}u$ blow up while the solution $u$ keeps to be small. For the 1D scalar equation or $2\times 2$ strictly hyperbolic system (corresponding to $n=1, 2$), if the smooth solution $u$ blows up in finite time, then the blowup mechanism can be well understood (i.e., only the blowup of $\partial_{t,x}u$ happens). In the present paper, for the $n\times n$ ($n\geq 3$) strictly hyperbolic system with a class of large initial data, we are concerned with the blowup mechanism of smooth solution $u$ on the finite blowup time and the detailed singularity behaviours of $\partial_{t,x}u$ near the blowup point. Our results are based on the efficient decomposition of $u$ along the different characteristic directions, the suitable introduction of the modulated coordinates and the global weighted energy estimates.

math.AP↗

arXiv2022

Optical Remote Sensing Image Understanding with Weak Supervision: Concepts, Methods, and Perspectives

In recent years, supervised learning has been widely used in various tasks of optical remote sensing image understanding, including remote sensing image classification, pixel-wise segmentation, change detection, and object detection. The methods based on supervised learning need a large amount of high-quality training data and their performance highly depends on the quality of the labels. However, in practical remote sensing applications, it is often expensive and time-consuming to obtain large-scale data sets with high-quality labels, which leads to a lack of sufficient supervised information. In some cases, only coarse-grained labels can be obtained, resulting in the lack of exact supervision. In addition, the supervised information obtained manually may be wrong, resulting in a lack of accurate supervision. Therefore, remote sensing image understanding often faces the problems of incomplete, inexact, and inaccurate supervised information, which will affect the breadth and depth of remote sensing applications. In order to solve the above-mentioned problems, researchers have explored various tasks in remote sensing image understanding under weak supervision. This paper summarizes the research progress of weakly supervised learning in the field of remote sensing, including three typical weakly supervised paradigms: 1) Incomplete supervision, where only a subset of training data is labeled; 2) Inexact supervision, where only coarse-grained labels of training data are given; 3) Inaccurate supervision, where the labels given are not always true on the ground.

cs.CV↗

arXiv2022

Performance Analysis of Wireless Network Aided by Discrete-Phase-Shifter IRS

Discrete phase shifters of intelligent reflecting surface (IRS) generates phase quantization error (QE) and degrades the receive performance at the receiver. To make an analysis of the performance loss caused by IRS with phase QE, based on the law of large numbers, the closed-form expressions of signal-to-noise ratio (SNR) performance loss (PL), achievable rate (AR), and bit error rate (BER) are successively derived under line-of-sight (LoS) channels and Rayleigh channels. Moreover, based on the Taylor series expansion, the approximate simple closed form of PL of IRS with approximate QE is also given. The simulation results show that the performance losses of SNR and AR decrease as the number of quantization bits increase, while they gradually increase with the number of IRS phase shifter elements increase. Regardless of LoS channels or Rayleigh channels, when the number of quantization bits is larger than or equal to 3, the performance losses of SNR and AR are less than 0.23dB and 0.08bits/s/Hz, respectively, and the BER performance degradation is trivial. In particular, the performance loss difference between IRS with QE and IRS with approximate QE is negligible when the number of quantization bits is not less than 2.

cs.IT↗

arXiv2022

Elevation Angle-Dependent 3D Trajectory Design for Aerial RIS-aided Communication

This paper investigates an aerial reconfigurable intelligent surface (RIS)-aided communication system under the probabilistic line-of-sight (LoS) channel, where an unmanned aerial vehicle (UAV) equipped with an RIS is deployed to assist two ground nodes in their information exchange. An optimization problem with the objective of maximizing the minimum average achievable rate is formulated to jointly design the communication scheduling, the RIS's phase shift, and the three-dimensional (3D) UAV trajectory. To solve such a non-convex problem, we propose an efficient iterative algorithm to obtain its suboptimal solution. Simulation results show that our proposed design significantly outperforms the existing schemes and provides new insights into the elevation angle and distance trade-off for the UAV-borne RIS communication system.

eess.SP↗

arXiv2022

A3CLNN: Spatial, Spectral and Multiscale Attention ConvLSTM Neural Network for Multisource Remote Sensing Data Classification

The problem of effectively exploiting the information multiple data sources has become a relevant but challenging research topic in remote sensing. In this paper, we propose a new approach to exploit the complementarity of two data sources: hyperspectral images (HSIs) and light detection and ranging (LiDAR) data. Specifically, we develop a new dual-channel spatial, spectral and multiscale attention convolutional long short-term memory neural network (called dual-channel A3CLNN) for feature extraction and classification of multisource remote sensing data. Spatial, spectral and multiscale attention mechanisms are first designed for HSI and LiDAR data in order to learn spectral- and spatial-enhanced feature representations, and to represent multiscale information for different classes. In the designed fusion network, a novel composite attention learning mechanism (combined with a three-level fusion strategy) is used to fully integrate the features in these two data sources. Finally, inspired by the idea of transfer learning, a novel stepwise training strategy is designed to yield a final classification result. Our experimental results, conducted on several multisource remote sensing data sets, demonstrate that the newly proposed dual-channel A3CLNN exhibits better feature representation ability (leading to more competitive classification performance) than other state-of-the-art methods.

cs.CV↗

arXiv2022

MC-UNet Multi-module Concatenation based on U-shape Network for Retinal Blood Vessels Segmentation

Accurate segmentation of the blood vessels of the retina is an important step in clinical diagnosis of ophthalmic diseases. Many deep learning frameworks have come up for retinal blood vessels segmentation tasks. However, the complex vascular structure and uncertain pathological features make the blood vessel segmentation still very challenging. A novel U-shaped network named Multi-module Concatenation which is based on Atrous convolution and multi-kernel pooling is put forward to retinal vessels segmentation in this paper. The proposed network structure retains three layers the essential structure of U-Net, in which the atrous convolution combining the multi-kernel pooling blocks are designed to obtain more contextual information. The spatial attention module is concatenated with dense atrous convolution module and multi-kernel pooling module to form a multi-module concatenation. And different dilation rates are selected by cascading to acquire a larger receptive field in atrous convolution. Adequate comparative experiments are conducted on these public retinal datasets: DRIVE, STARE and CHASE_DB1. The results show that the proposed method is effective, especially for microvessels. The code will be put out at https://github.com/Rebeccala/MC-UNet

eess.IV↗

arXiv2022

Semi-Data-Aided Channel Estimation for MIMO Systems via Reinforcement Learning

Data-aided channel estimation is a promising solution to improve channel estimation accuracy by exploiting data symbols as pilot signals for updating an initial channel estimate. In this paper, we propose a semi-data-aided channel estimator for multiple-input multiple-output communication systems. Our strategy is to leverage reinforcement learning (RL) for selecting reliable detected symbols among the symbols in the first part of transmitted data block. This strategy facilitates an update of the channel estimate before the end of data block transmission and therefore achieves a significant reduction in communication latency compared to conventional data-aided channel estimation approaches. Towards this end, we first define a Markov decision process (MDP) which sequentially decides whether to use each detected symbol as an additional pilot signal. We then develop an RL algorithm to efficiently find the best policy of the MDP based on a Monte Carlo tree search approach. In this algorithm, we exploit the a-posteriori probability for approximating both the optimal future actions and the corresponding state transitions of the MDP and derive a closed-form expression for the best policy. Simulation results demonstrate that the proposed channel estimator effectively mitigates both channel estimation error and detection performance loss caused by insufficient pilot signals.

eess.SP↗

arXiv2022

Energy-Efficient IRS-Aided NOMA Beamforming for 6G Wireless Communications

This manuscript presents an energy-efficient alternating optimization framework based on intelligent reflective surfaces (IRS) aided non-orthogonal multiple access beamforming (NOMA-BF) system for 6G wireless communications. Specifically, this work proposes a centralized IRS-enabled design for the NOMA-BF system to optimize the active beamforming and power allocation coefficient (PAC) of users at the transmitter in the first stage and passive beamforming at IRS in the 2nd stage to maximize the energy efficiency (EE) of the network. However, an increment in the number of supportable users with the NOMA-BF system will lead to NOMA user interference and inter-cluster interference (ICI). To mitigate the effect of ICI, first zero-forcing beamforming along with efficient user clustering algorithm is exploited and then NOMA user interference is tackled efficiently through a proposed iterative algorithm that computes PAC of NOMA user through simplified closed-form expression under the required system constraints. In the 2nd stage, the problem of passive beamforming is solved through a technique based on difference-of-convex (DC) programming and successive convex approximation (SCA). Simulation results demonstrate that the proposed alternating framework for energy-efficient IRS-assisted NOMA-BF system can achieve convergence within a few iterations and provide efficient performance in terms of EE of the system with low complexity.

eess.SP↗

arXiv2022

Beamforming Design for IRS-Aided Decode-and-Forward Relay Wireless Network

As a low-cost and low-power-consumption passive reflector, intelligent reflecting surface (IRS) can make a significant rate improvement by building a programmable wireless environment. To improve the rate performance and coverage range of wireless networks, an IRS-aided decode-and-forward (DF) relay network is proposed with multiple antennas at relay station (RS). To achieve a high rate, an alternately iterative structure (AIS) of maximizing receive power (Max-RP) at RS is proposed to jointly optimize the beamforming vectors at RS and phase shifts at IRS. Considering its high-complexity, two low-complexity Max-RP schemes of null-space projection (NSP) plus maximum ratio combining (MRC) and IRS element selection (IRSES) plus MRC are presented to reduce this complexity, respectively. For the former, NSP is used to separate the reflected signal from IRS and the direct transmitted signal from source and MRC is adopted to combine the two signals at RS. For the latter, the basic concept of IRSES is as follows: IRS is partitioned into M subsets of elements and adjusting the phases of all elements per subset make all reflected signals and the direct signal from source phase alignment (PA) at the corresponding antenna of relay. Simulation results show that the proposed three methods perform much better than the existing network with single-antenna relay in terms of rate performance. In particular, a 85% rate gain over existing scheme is achieved in the high signal-to-noise ratio region. Moreover, it is verified that the positions of RS and IRS have a substantial impact on rate performance, and there exists an optimal positions of RS and IRS.

cs.IT↗

arXiv2022

Collaborative Intelligent Reflecting Surface Networks with Multi-Agent Reinforcement Learning

Intelligent reflecting surface (IRS) is envisioned to be widely applied in future wireless networks. In this paper, we investigate a multi-user communication system assisted by cooperative IRS devices with the capability of energy harvesting. Aiming to maximize the long-term average achievable system rate, an optimization problem is formulated by jointly designing the transmit beamforming at the base station (BS) and discrete phase shift beamforming at the IRSs, with the constraints on transmit power, user data rate requirement and IRS energy buffer size. Considering time-varying channels and stochastic arrivals of energy harvested by the IRSs, we first formulate the problem as a Markov decision process (MDP) and then develop a novel multi-agent Q-mix (MAQ) framework with two layers to decouple the optimization parameters. The higher layer is for optimizing phase shift resolutions, and the lower one is for phase shift beamforming and power allocation. Since the phase shift optimization is an integer programming problem with a large-scale action space, we improve MAQ by incorporating the Wolpertinger method, namely, MAQ-WP algorithm to achieve a sub-optimality with reduced dimensions of action space. In addition, as MAQ-WP is still of high complexity to achieve good performance, we propose a policy gradient-based MAQ algorithm, namely, MAQ-PG, by mapping the discrete phase shift actions into a continuous space at the cost of a slight performance loss. Simulation results demonstrate that the proposed MAQ-WP and MAQ-PG algorithms can converge faster and achieve data rate improvements of 10.7% and 8.8% over the conventional multi-agent DDPG, respectively.

cs.MA↗

arXiv2022

Industrial Style Transfer with Large-scale Geometric Warping and Content Preservation

We propose a novel style transfer method to quickly create a new visual product with a nice appearance for industrial designers' reference. Given a source product, a target product, and an art style image, our method produces a neural warping field that warps the source shape to imitate the geometric style of the target and a neural texture transformation network that transfers the artistic style to the warped source product. Our model, Industrial Style Transfer (InST), consists of large-scale geometric warping (LGW) and interest-consistency texture transfer (ICTT). LGW aims to explore an unsupervised transformation between the shape masks of the source and target products for fitting large-scale shape warping. Furthermore, we introduce a mask smoothness regularization term to prevent the abrupt changes of the details of the source product. ICTT introduces an interest regularization term to maintain important contents of the warped product when it is stylized by using the art style image. Extensive experimental results demonstrate that InST achieves state-of-the-art performance on multiple visual product design tasks, e.g., companies' snail logos and classical bottles (please see Fig. 1). To the best of our knowledge, we are the first to extend the neural style transfer method to create industrial product appearances. Project page: \ulr{https://jcyang98.github.io/InST/home.html}. Code available at: \url{https://github.com/jcyang98/InST}.

cs.CV↗

arXiv2022

Finite Sample t-Tests for High-Dimensional Means

Size distortion can occur if an asymptotic testing procedure requiring diverging sample sizes, is implemented to data with very small sample sizes. In this paper, we consider one-sample and two-sample tests for mean vectors when data are high-dimensional but sample sizes are very small. We establish asymptotic t-distributions of one-sample and two-sample U-statistics, which only require data dimensionality to diverge but sample sizes to be fixed and no less than 3. Simulation studies confirm the theoretical results that the proposed tests maintain accurate empirical sizes for a wide range of sample sizes and data dimensionalities. We apply the proposed tests to an fMRI dataset to demonstrate the practical implementation of the methods.

stat.ME↗

arXiv2022

Universal Segmentation of 33 Anatomies

In the paper, we present an approach for learning a single model that universally segments 33 anatomical structures, including vertebrae, pelvic bones, and abdominal organs. Our model building has to address the following challenges. Firstly, while it is ideal to learn such a model from a large-scale, fully-annotated dataset, it is practically hard to curate such a dataset. Thus, we resort to learn from a union of multiple datasets, with each dataset containing the images that are partially labeled. Secondly, along the line of partial labelling, we contribute an open-source, large-scale vertebra segmentation dataset for the benefit of spine analysis community, CTSpine1K, boasting over 1,000 3D volumes and over 11K annotated vertebrae. Thirdly, in a 3D medical image segmentation task, due to the limitation of GPU memory, we always train a model using cropped patches as inputs instead a whole 3D volume, which limits the amount of contextual information to be learned. To this, we propose a cross-patch transformer module to fuse more information in adjacent patches, which enlarges the aggregated receptive field for improved segmentation performance. This is especially important for segmenting, say, the elongated spine. Based on 7 partially labeled datasets that collectively contain about 2,800 3D volumes, we successfully learn such a universal model. Finally, we evaluate the universal model on multiple open-source datasets, proving that our model has a good generalization performance and can potentially serve as a solid foundation for downstream tasks.

eess.IV↗

arXiv2022

MixCL: Pixel label matters to contrastive learning

Contrastive learning and self-supervised techniques have gained prevalence in computer vision for the past few years. It is essential for medical image analysis, which is often notorious for its lack of annotations. Most existing self-supervised methods applied in natural imaging tasks focus on designing proxy tasks for unlabeled data. For example, contrastive learning is often based on the fact that an image and its transformed version share the same identity. However, pixel annotations contain much valuable information for medical image segmentation, which is largely ignored in contrastive learning. In this work, we propose a novel pre-training framework called Mixed Contrastive Learning (MixCL) that leverages both image identities and pixel labels for better modeling by maintaining identity consistency, label consistency, and reconstruction consistency together. Consequently, thus pre-trained model has more robust representations that characterize medical images. Extensive experiments demonstrate the effectiveness of the proposed method, improving the baseline by 5.28% and 14.12% in Dice coefficient when 5% labeled data of Spleen and 15% of BTVC are used in fine-tuning, respectively.

eess.IV↗