arXiv subjects

Jing Zhang

Publications and source records attributed to Jing Zhang.

At least 649 records · Page 36Linked to original sources

arXiv2022

RegionCL: Can Simple Region Swapping Contribute to Contrastive Learning?

Self-supervised methods (SSL) have achieved significant success via maximizing the mutual information between two augmented views, where cropping is a popular augmentation technique. Cropped regions are widely used to construct positive pairs, while the left regions after cropping have rarely been explored in existing methods, although they together constitute the same image instance and both contribute to the description of the category. In this paper, we make the first attempt to demonstrate the importance of both regions in cropping from a complete perspective and propose a simple yet effective pretext task called Region Contrastive Learning (RegionCL). Specifically, given two different images, we randomly crop a region (called the paste view) from each image with the same size and swap them to compose two new images together with the left regions (called the canvas view), respectively. Then, contrastive pairs can be efficiently constructed according to the following simple criteria, i.e., each view is (1) positive with views augmented from the same original image and (2) negative with views augmented from other images. With minor modifications to popular SSL methods, RegionCL exploits those abundant pairs and helps the model distinguish the regions features from both canvas and paste views, therefore learning better visual representations. Experiments on ImageNet, MS COCO, and Cityscapes demonstrate that RegionCL improves MoCo v2, DenseCL, and SimSiam by large margins and achieves state-of-the-art performance on classification, detection, and segmentation tasks. The code will be available at https://github.com/Annbless/RegionCL.git.

cs.CV↗

arXiv2022

Rethinking Hierarchies in Pre-trained Plain Vision Transformer

Self-supervised pre-training vision transformer (ViT) via masked image modeling (MIM) has been proven very effective. However, customized algorithms should be carefully designed for the hierarchical ViTs, e.g., GreenMIM, instead of using the vanilla and simple MAE for the plain ViT. More importantly, since these hierarchical ViTs cannot reuse the off-the-shelf pre-trained weights of the plain ViTs, the requirement of pre-training them leads to a massive amount of computational cost, thereby incurring both algorithmic and computational complexity. In this paper, we address this problem by proposing a novel idea of disentangling the hierarchical architecture design from the self-supervised pre-training. We transform the plain ViT into a hierarchical one with minimal changes. Technically, we change the stride of linear embedding layer from 16 to 4 and add convolution (or simple average) pooling layers between the transformer blocks, thereby reducing the feature size from 1/4 to 1/32 sequentially. Despite its simplicity, it outperforms the plain ViT baseline in classification, detection, and segmentation tasks on ImageNet, MS COCO, Cityscapes, and ADE20K benchmarks, respectively. We hope this preliminary study could draw more attention from the community on developing effective (hierarchical) ViTs while avoiding the pre-training cost by leveraging the off-the-shelf checkpoints. The code and models will be released at https://github.com/ViTAE-Transformer/HPViT.

cs.CV↗

arXiv2022

Watermarking for Out-of-distribution Detection

Out-of-distribution (OOD) detection aims to identify OOD data based on representations extracted from well-trained deep models. However, existing methods largely ignore the reprogramming property of deep models and thus may not fully unleash their intrinsic strength: without modifying parameters of a well-trained deep model, we can reprogram this model for a new purpose via data-level manipulation (e.g., adding a specific feature perturbation to the data). This property motivates us to reprogram a classification model to excel at OOD detection (a new task), and thus we propose a general methodology named watermarking in this paper. Specifically, we learn a unified pattern that is superimposed onto features of original data, and the model's detection capability is largely boosted after watermarking. Extensive experiments verify the effectiveness of watermarking, demonstrating the significance of the reprogramming property of deep models in OOD detection.

cs.LG↗

arXiv2022

Realization of higher Precision in Interferometric-weak-value-based Small-tilt Measurement

We experimentally realize a great precision enhancement in the small tilt measurement by using a Sagnac interferometer and balanced homodyne detection (BHD) of high-order optical modes, together with the weak value amplification (WVA) technique. Smaller minimum measurable tilt (MMT) and higher signal-to-noise ratio (SNR) can be obtained by using BHD, compared with the split detection (SD). The precision of 3.8 nrad can be obtained under our present experimental condition. It is shown that combining WVA technique and BHD can strengthen each other's advantages and can behave better for some special application scenarios, such as extremely weak output, wider measurement bandwidth, etc. Moreover, the precision can be further enhanced by experimental parameter optimization.

quant-ph↗

arXiv2022

Inference and Prediction Using Functional Principal Components Analysis: Application to Diabetic Kidney Disease Progression in the Chronic Renal Insufficiency Cohort (CRIC) Study

Repeated longitudinal measurements are commonly used to model long-term disease progression, and timing and number of assessments per patient may vary, leading to irregularly spaced and sparse data. Longitudinal trajectories may exhibit curvilinear patterns, in which mixed linear regression methods may fail to capture true trends in the data. We applied functional principal components analysis to model kidney disease progression via estimated glomerular filtration rate (eGFR) trajectories. In a cohort of 2641 participants with diabetes and up to 15 years of annual follow-up from the Chronic Renal Insufficiency Cohort (CRIC) study, we detected novel dominant modes of variation and patterns of diabetic kidney disease (DKD) progression among subgroups defined by the presence of albuminuria. We conducted inferential permutation tests to assess differences in longitudinal eGFR patterns between groups. To determine whether fitting a full cohort model or separate group-specific models is more optimal for modeling long-term trajectories, we evaluated model fit, using our goodness-of-fit procedure, and future prediction accuracy. Our findings indicated advantages for both modeling approaches in accomplishing different objectives. Beyond DKD, the methods described are applicable to other settings with longitudinally assessed biomarkers as indicators of disease progression. Supplementary materials for this article are available online.

stat.AP↗

arXiv2022

Oscillatory cooperation prevalence emerges from misperception

Oscillatory behaviors are ubiquitous in nature and the human society. However, most previous works fail to reproduce them in the two-strategy game-theoretical models. Here we show that oscillatory behaviors naturally emerge if incomplete information is incorporated into the cooperation evolution of a non-Markov model. Specifically, we consider a population playing prisoner's dilemma game, where each individual can only probabilistically get access to their neighbors' payoff information and store them within their memory with a given length. They make their decisions based upon these memories. Interestingly, we find that the level of cooperation generally cannot stabilize but render quasi-periodic oscillation, and this observation is strengthened for a longer memory and a smaller information acquisition probability. The mechanism uncovered shows that there are misperceived payoffs about the player's neighborhood, facilitating the growth of cooperators and defectors at different stages that leads to oscillatory behaviors as a result. Our findings are robust to the underlying structure of the population. Given the omnipresence of incomplete information, our findings may provide a plausible explanation for the phenomenon of oscillatory behaviors in the real world.

q-bio.PE↗

arXiv2022

ViTPose: Simple Vision Transformer Baselines for Human Pose Estimation

Although no specific domain knowledge is considered in the design, plain vision transformers have shown excellent performance in visual recognition tasks. However, little effort has been made to reveal the potential of such simple structures for pose estimation tasks. In this paper, we show the surprisingly good capabilities of plain vision transformers for pose estimation from various aspects, namely simplicity in model structure, scalability in model size, flexibility in training paradigm, and transferability of knowledge between models, through a simple baseline model called ViTPose. Specifically, ViTPose employs plain and non-hierarchical vision transformers as backbones to extract features for a given person instance and a lightweight decoder for pose estimation. It can be scaled up from 100M to 1B parameters by taking the advantages of the scalable model capacity and high parallelism of transformers, setting a new Pareto front between throughput and performance. Besides, ViTPose is very flexible regarding the attention type, input resolution, pre-training and finetuning strategy, as well as dealing with multiple pose tasks. We also empirically demonstrate that the knowledge of large ViTPose models can be easily transferred to small ones via a simple knowledge token. Experimental results show that our basic ViTPose model outperforms representative methods on the challenging MS COCO Keypoint Detection benchmark, while the largest model sets a new state-of-the-art. The code and models are available at https://github.com/ViTAE-Transformer/ViTPose.

cs.CV↗

arXiv2022

APT-36K: A Large-scale Benchmark for Animal Pose Estimation and Tracking

Animal pose estimation and tracking (APT) is a fundamental task for detecting and tracking animal keypoints from a sequence of video frames. Previous animal-related datasets focus either on animal tracking or single-frame animal pose estimation, and never on both aspects. The lack of APT datasets hinders the development and evaluation of video-based animal pose estimation and tracking methods, limiting real-world applications, e.g., understanding animal behavior in wildlife conservation. To fill this gap, we make the first step and propose APT-36K, i.e., the first large-scale benchmark for animal pose estimation and tracking. Specifically, APT-36K consists of 2,400 video clips collected and filtered from 30 animal species with 15 frames for each video, resulting in 36,000 frames in total. After manual annotation and careful double-check, high-quality keypoint and tracking annotations are provided for all the animal instances. Based on APT-36K, we benchmark several representative models on the following three tracks: (1) supervised animal pose estimation on a single frame under intra- and inter-domain transfer learning settings, (2) inter-species domain generalization test for unseen animals, and (3) animal pose estimation with animal tracking. Based on the experimental results, we gain some empirical insights and show that APT-36K provides a valuable animal pose estimation and tracking benchmark, offering new challenges and opportunities for future research. The code and dataset will be made publicly available at https://github.com/pandorgan/APT-36K.

cs.CV↗

arXiv2022

A Comprehensive Survey on Data-Efficient GANs in Image Generation

Generative Adversarial Networks (GANs) have achieved remarkable achievements in image synthesis. These successes of GANs rely on large scale datasets, requiring too much cost. With limited training data, how to stable the training process of GANs and generate realistic images have attracted more attention. The challenges of Data-Efficient GANs (DE-GANs) mainly arise from three aspects: (i) Mismatch Between Training and Target Distributions, (ii) Overfitting of the Discriminator, and (iii) Imbalance Between Latent and Data Spaces. Although many augmentation and pre-training strategies have been proposed to alleviate these issues, there lacks a systematic survey to summarize the properties, challenges, and solutions of DE-GANs. In this paper, we revisit and define DE-GANs from the perspective of distribution optimization. We conclude and analyze the challenges of DE-GANs. Meanwhile, we propose a taxonomy, which classifies the existing methods into three categories: Data Selection, GANs Optimization, and Knowledge Sharing. Last but not the least, we attempt to highlight the current problems and the future directions.

cs.CV↗

arXiv2022

The boundedness of commutators of sublinear operators on Herz Triebel-Lizorkin spaces with variable exponent

In this paper, the authors first discuss the characterization of Herz Triebel-Lizorkin spaces with variable exponent via two families of operators. By this characterization, the authors prove that the Lipschitz commutators of sublinear operators is bounded from Herz spaces with variable exponent to Herz Triebel-Lizorkin spaces with variable exponent. As an application, the corresponding boundedness estimates for the commutators of maximal operator, Riesz potential operator and Calderón-Zygmund operator are established.

math.FA↗

arXiv2022

A General Divergence Modeling Strategy for Salient Object Detection

Salient object detection is subjective in nature, which implies that multiple estimations should be related to the same input image. Most existing salient object detection models are deterministic following a point to point estimation learning pipeline, making them incapable of estimating the predictive distribution. Although latent variable model based stochastic prediction networks exist to model the prediction variants, the latent space based on the single clean saliency annotation is less reliable in exploring the subjective nature of saliency, leading to less effective saliency divergence modeling. Given multiple saliency annotations, we introduce a general divergence modeling strategy via random sampling, and apply our strategy to an ensemble based framework and three latent variable model based solutions to explore the subjective nature of saliency. Experimental results prove the superior performance of our general divergence modeling strategy.

cs.CV↗

arXiv2022

Decentralized Coordinated Precoding Design in Cell-Free Massive MIMO Systems for URLLC

Cell-free massive multiple-input multiple-output (MIMO) is a promising network to offer huge improvement of the achievable rate compared with conventional cellular massive MIMO systems. However, the commonly adopted Shannon-type achievable rate is only valid in the long block length regime that is not applicable to the emerging short-packet communication. To realize ultra-reliable and low-latency communication (URLLC) in cell-free massive MIMO systems, we optimize the precoding vector at the access points (APs) to maximize the minimum user rate in both the centralized and decentralized fashion. The design takes into account the impact of URLLC and we propose path-following algorithms (PFA) to address the considered problem which generates a sequence of advanced feasible points and converges to at least a locally optimal solution of the design problem. Moreover, we investigate the requirement of the precoding schemes, the length of the transmission duration, the number of antennas equipped at each AP, and the size of each AP cluster on the URLLC rate. Numerical results show that the decentralized PFA precoding can achieve 80\% of the 95\%-likely URLLC rate of the centralized precoding and 89\% of the average URLLC rate with only 12\% computational complexity of the centralized precoding.

cs.IT↗

arXiv2022

Quantum Hall phase in graphene engineered by interfacial charge coupling

Quantum Hall effect (QHE), the ground to construct modern conceptual electronic systems with emerging physics, is often much influenced by the interplay between the host two-dimensional electron gases and the substrate, sometimes predicted to exhibit exotic topological states. Yet the understanding of the underlying physics and the controllable engineering of this paradigm of interaction remain challenging. Here we demonstrate the observation of an unusual QHE, which differs markedly from the known picture, in graphene samples in contact with an anti-ferromagnetic insulator CrOCl equipped with dual gates. Owing to the peculiar interfacial coupling, Landau levels in monolayer graphene remain intact at negative filling fractions, but largely deviated for the positive gate-doping range. The latter QHE phase even presents in the limit of zero magnetic field, with the consequential Landau quantization following a parabolic relation between the displacement field $D$ and the magnetic field $B$. This characteristic prevails up to 100 K in a sufficiently wide effective doping range from 0 to 10$^{13}$ cm$^{-2}$. Our findings thus open up new routes for manipulating the quantum electronic states, which may find applications in such as quantum metrology.

cond-mat.mes-hall↗

arXiv2022

Towards Explainable 3D Grounded Visual Question Answering: A New Benchmark and Strong Baseline

Recently, 3D vision-and-language tasks have attracted increasing research interest. Compared to other vision-and-language tasks, the 3D visual question answering (VQA) task is less exploited and is more susceptible to language priors and co-reference ambiguity. Meanwhile, a couple of recently proposed 3D VQA datasets do not well support 3D VQA task due to their limited scale and annotation methods. In this work, we formally define and address a 3D grounded VQA task by collecting a new 3D VQA dataset, referred to as FE-3DGQA, with diverse and relatively free-form question-answer pairs, as well as dense and completely grounded bounding box annotations. To achieve more explainable answers, we labelled the objects appeared in the complex QA pairs with different semantic types, including answer-grounded objects (both appeared and not appeared in the questions), and contextual objects for answer-grounded objects. We also propose a new 3D VQA framework to effectively predict the completely visually grounded and explainable answer. Extensive experiments verify that our newly collected benchmark datasets can be effectively used to evaluate various 3D VQA methods from different aspects and our newly proposed framework also achieves state-of-the-art performance on the new benchmark dataset. Both the newly collected dataset and our codes will be publicly available at http://github.com/zlccccc/3DGQA.

cs.CV↗

arXiv2022

Approximating diamond principles on products at an inaccessible cardinal

We isolate \emph{the approximating diamond principles}, which are consequences of the diamond principle at an inaccessible cardinal. We use these principles to find new methods for negating the diamond principle at large cardinals. Most notably, we demonstrate, using Gitik's overlapping extenders forcing, a new method to get the consistency of the failure of the diamond principle at a large cardinal $θ$ without changing cofinalities or adding fast clubs to $θ$. In addition, we show that the approximating diamond principles necessarily hold at a weakly compact cardinal. This result, combined with the fact that in all known models where the diamond principle fails the approximating diamond principles also fail at an inaccessible cardinal, exhibits essential combinatorial obstacles to make the diamond principle fail at a weakly compact cardinal.

math.LO↗

arXiv2022

Tuning of Quantum Paraelectricity of M-type Hexaferrite BaFe12O19 by External Parameters

M-type hexaferrite BaFe12O19 was recently reported to be a new type of quantum paraelectrics with triangular lattice by showing a low temperature dielectric plateau due to quantum fluctuation. It has also been proposed to have a possible quantum-dipole liquid ground state. To suppress its quantum fluctuations and reach a possible quantum critical point, we have tuned its quantum paraelectricity in three ways: (i) 57Fe isotope replacement; (ii) in-plane compressive strain; and (iii) hydrostatic pressure. It is found that 95% 57Fe replacement and the in-plane strain are more effective to drive its ground state closer to a critical region by inducing a peak feature in the temperature dependence of dielectric constant. In contrast, the application of hydrostatic pressure pushed the system away from the quantum critical point by gradually suppressing the plateau feature in dielectric constant. Our combined efforts reveal the potential of the M-type hexaferrites for studying the quantum critical behaviors.

cond-mat.str-el↗

arXiv2022

Improving RGB-D Point Cloud Registration by Learning Multi-scale Local Linear Transformation

Point cloud registration aims at estimating the geometric transformation between two point cloud scans, in which point-wise correspondence estimation is the key to its success. In addition to previous methods that seek correspondences by hand-crafted or learnt geometric features, recent point cloud registration methods have tried to apply RGB-D data to achieve more accurate correspondence. However, it is not trivial to effectively fuse the geometric and visual information from these two distinctive modalities, especially for the registration problem. In this work, we propose a new Geometry-Aware Visual Feature Extractor (GAVE) that employs multi-scale local linear transformation to progressively fuse these two modalities, where the geometric features from the depth data act as the geometry-dependent convolution kernels to transform the visual features from the RGB data. The resultant visual-geometric features are in canonical feature spaces with alleviated visual dissimilarity caused by geometric changes, by which more reliable correspondence can be achieved. The proposed GAVE module can be readily plugged into recent RGB-D point cloud registration framework. Extensive experiments on 3D Match and ScanNet demonstrate that our method outperforms the state-of-the-art point cloud registration methods even without correspondence or pose supervision. The code is available at: https://github.com/514DNA/LLT.

cs.CV↗

arXiv2022

Graph Contrastive Learning for Anomaly Detection

Graph-based anomaly detection has been widely used for detecting malicious activities in real-world applications. Existing attempts to address this problem have thus far focused on structural feature engineering or learning in the binary classification regime. In this work, we propose to leverage graph contrastive coding and present the supervised GraphCAD model for contrasting abnormal nodes with normal ones in terms of their distances to the global context (e.g., the average of all nodes). To handle scenarios with scarce labels, we further enable GraphCAD as a self-supervised framework by designing a graph corrupting strategy for generating synthetic node labels. To achieve the contrastive objective, we design a graph neural network encoder that can infer and further remove suspicious links during message passing, as well as learn the global context of the input graph. We conduct extensive experiments on four public datasets, demonstrating that 1) GraphCAD significantly and consistently outperforms various advanced baselines and 2) its self-supervised version without fine-tuning can achieve comparable performance with its fully supervised version.

cs.LG↗