arXiv subjects

Wei Zhao

Publications and source records attributed to Wei Zhao.

At least 307 records · Page 17Linked to original sources

arXiv2021

TransCT: Dual-path Transformer for Low Dose Computed Tomography

Low dose computed tomography (LDCT) has attracted more and more attention in routine clinical diagnosis assessment, therapy planning, etc., which can reduce the dose of X-ray radiation to patients. However, the noise caused by low X-ray exposure degrades the CT image quality and then affects clinical diagnosis accuracy. In this paper, we train a transformer-based neural network to enhance the final CT image quality. To be specific, we first decompose the noisy LDCT image into two parts: high-frequency (HF) and low-frequency (LF) compositions. Then, we extract content features (X_{L_c}) and latent texture features (X_{L_t}) from the LF part, as well as HF embeddings (X_{H_f}) from the HF part. Further, we feed X_{L_t} and X_{H_f} into a modified transformer with three encoders and decoders to obtain well-refined HF texture features. After that, we combine these well-refined HF texture features with the pre-extracted X_{L_c} to encourage the restoration of high-quality LDCT images with the assistance of piecewise reconstruction. Extensive experiments on Mayo LDCT dataset show that our method produces superior results and outperforms other methods.

eess.IV↗

arXiv2021

On the geometry of irreversible metric-measure spaces: Convergence, Stability and Analytic aspects

The paper is devoted to the study of Gromov-Hausdorff convergence and stability of irreversible metric-measure spaces, both in the compact and noncompact cases. While the compact setting is mostly similar to the reversible case developed by J. Lott, K.-T. Sturm and C. Villani, the noncompact case provides various surprising phenomena. Since the reversibility of noncompact irreversible spaces might be infinite, it is motivated to introduce a suitable nondecreasing function that bounds the reversibility of larger and larger balls. By this approach, we are able to prove satisfactory convergence/stability results in a suitable -- reversibility depending -- Gromov-Hausdorff topology. A wide class of irreversible spaces is provided by Finsler manifolds, which serve to construct various model examples by pointing out genuine differences between the reversible and irreversible settings. We conclude the paper by proving various geometric and functional inequalities (as Brunn-Minkowski, Bishop-Gromov, log-Sobolev and Lichnerowicz inequalities) on irreversible structures.

math.DG↗

arXiv2021

Graph Routing between Capsules

Routing methods in capsule networks often learn a hierarchical relationship for capsules in successive layers, but the intra-relation between capsules in the same layer is less studied, while this intra-relation is a key factor for the semantic understanding in text data. Therefore, in this paper, we introduce a new capsule network with graph routing to learn both relationships, where capsules in each layer are treated as the nodes of a graph. We investigate strategies to yield adjacency and degree matrix with three different distances from a layer of capsules, and propose the graph routing mechanism between those capsules. We validate our approach on five text classification datasets, and our findings suggest that the approach combining bottom-up routing and top-down attention performs the best. Such an approach demonstrates generalization capability across datasets. Compared to the state-of-the-art routing methods, the improvements in accuracy in the five datasets we used were 0.82, 0.39, 0.07, 1.01, and 0.02, respectively.

cs.LG↗

arXiv2021

Inducing Language-Agnostic Multilingual Representations

Cross-lingual representations have the potential to make NLP techniques available to the vast majority of languages in the world. However, they currently require large pretraining corpora or access to typologically similar languages. In this work, we address these obstacles by removing language identity signals from multilingual embeddings. We examine three approaches for this: (i) re-aligning the vector spaces of target languages (all together) to a pivot source language; (ii) removing language-specific means and variances, which yields better discriminativeness of embeddings as a by-product; and (iii) increasing input similarity across languages by removing morphological contractions and sentence reordering. We evaluate on XNLI and reference-free MT across 19 typologically diverse languages. Our findings expose the limitations of these approaches -- unlike vector normalization, vector space re-alignment and text normalization do not achieve consistent gains across encoders and languages. Due to the approaches' additive effects, their combination decreases the cross-lingual transfer gap by 8.9 points (m-BERT) and 18.2 points (XLM-R) on average across all tasks and languages, however. Our code and models are publicly available.

cs.CL↗

arXiv2021

Multi-Agent Cooperative Bidding Games for Multi-Objective Optimization in e-Commercial Sponsored Search

Bid optimization for online advertising from single advertiser's perspective has been thoroughly investigated in both academic research and industrial practice. However, existing work typically assume competitors do not change their bids, i.e., the wining price is fixed, leading to poor performance of the derived solution. Although a few studies use multi-agent reinforcement learning to set up a cooperative game, they still suffer the following drawbacks: (1) They fail to avoid collusion solutions where all the advertisers involved in an auction collude to bid an extremely low price on purpose. (2) Previous works cannot well handle the underlying complex bidding environment, leading to poor model convergence. This problem could be amplified when handling multiple objectives of advertisers which are practical demands but not considered by previous work. In this paper, we propose a novel multi-objective cooperative bid optimization formulation called Multi-Agent Cooperative bidding Games (MACG). MACG sets up a carefully designed multi-objective optimization framework where different objectives of advertisers are incorporated. A global objective to maximize the overall profit of all advertisements is added in order to encourage better cooperation and also to protect self-bidding advertisers. To avoid collusion, we also introduce an extra platform revenue constraint. We analyze the optimal functional form of the bidding formula theoretically and design a policy network accordingly to generate auction-level bids. Then we design an efficient multi-agent evolutionary strategy for model optimization. Offline experiments and online A/B tests conducted on the Taobao platform indicate both single advertiser's objective and global profit have been significantly improved compared to state-of-art methods.

cs.AI↗

arXiv2021

A Geometry-Informed Deep Learning Framework for Ultra-Sparse 3D Tomographic Image Reconstruction

Deep learning affords enormous opportunities to augment the armamentarium of biomedical imaging, albeit its design and implementation have potential flaws. Fundamentally, most deep learning models are driven entirely by data without consideration of any prior knowledge, which dramatically increases the complexity of neural networks and limits the application scope and model generalizability. Here we establish a geometry-informed deep learning framework for ultra-sparse 3D tomographic image reconstruction. We introduce a novel mechanism for integrating geometric priors of the imaging system. We demonstrate that the seamless inclusion of known priors is essential to enhance the performance of 3D volumetric computed tomography imaging with ultra-sparse sampling. The study opens new avenues for data-driven biomedical imaging and promises to provide substantially improved imaging tools for various clinical imaging and image-guided interventions.

cs.CV↗

arXiv2021

Importance Weighted Adversarial Discriminative Transfer for Anomaly Detection

Previous transfer methods for anomaly detection generally assume the availability of labeled data in source or target domains. However, such an assumption is not valid in most real applications where large-scale labeled data are too expensive. Therefore, this paper proposes an importance weighted adversarial autoencoder-based method to transfer anomaly detection knowledge in an unsupervised manner, particularly for a rarely studied scenario where a target domain has no labeled normal/abnormal data while only normal data from a related source domain exist. Specifically, the method learns to align the distributions of normal data in both source and target domains, but leave the distribution of abnormal data in the target domain unchanged. In this way, an obvious gap can be produced between the distributions of normal and abnormal data in the target domain, therefore enabling the anomaly detection in the domain. Extensive experiments on multiple synthetic datasets and the UCSD benchmark demonstrate the effectiveness of our approach.

cs.LG↗

arXiv2021

Contracting over persistent information

We consider a dynamic moral hazard problem between a principal and an agent, where the sole instrument the principal has to incentivize the agent is the disclosure of information. The principal aims at maximizing the (discounted) number of times the agent chooses a particular action, e.g., to work hard. We show that there exists an optimal contract, where the principal stops disclosing information as soon as its most preferred action is a static best reply for the agent or else continues disclosing information until the agent perfectly learns the principal's private information. If the agent perfectly learns the state, he learns it in finite time with probability one; the more patient the agent, the later he learns it.

econ.TH↗

arXiv2021

The Observation of Ferroelastic and Ferrielectric Domains in AgNbO3 Single Crystal

Compared to AgNbO3 based ceramics, the experimental investigations on the single crystalline AgNbO3, especially the ground state and ferroic domain structures, are not on the same level. Here in this work, based on successfully synthesized AgNbO3 single crystal using flux method, we observed the coexistence of ferroelastic and ferrielectric domain structures by a combination study of polarized light microscopy and piezoresponse force microscope, this finding may provide a new aspect for studying AgNbO3. The result also suggests a weak electromechanical response from the ferrielectric phase of AgNbO3 which is also supported by the transmission electron microscope characterization. Our results reveal that the AgNbO3 single crystal is in a polar ferrielectric phase at room temperature, clarifying its ground state which is controversial from the AgNbO3 ceramic materials.

cond-mat.mtrl-sci↗

arXiv2021

Structural Interventions in Networks

Two types of interventions are commonly implemented in networks: characteristic intervention, which influences individuals' intrinsic incentives, and structural intervention, which targets the social links among individuals. In this paper we provide a general framework to evaluate the distinct equilibrium effects of both types of interventions. We identify a hidden equivalence between a structural intervention and an endogenously determined characteristic intervention. Compared with existing approaches in the literature, the perspective from such an equivalence provides several advantages in the analysis of interventions that target network structure. We present a wide range of applications of our theory, including identifying the most wanted criminal(s) in delinquent networks and targeting the key connector for isolated communities.

econ.TH↗

arXiv2021

A novel multiple instance learning framework for COVID-19 severity assessment via data augmentation and self-supervised learning

How to fast and accurately assess the severity level of COVID-19 is an essential problem, when millions of people are suffering from the pandemic around the world. Currently, the chest CT is regarded as a popular and informative imaging tool for COVID-19 diagnosis. However, we observe that there are two issues -- weak annotation and insufficient data that may obstruct automatic COVID-19 severity assessment with CT images. To address these challenges, we propose a novel three-component method, i.e., 1) a deep multiple instance learning component with instance-level attention to jointly classify the bag and also weigh the instances, 2) a bag-level data augmentation component to generate virtual bags by reorganizing high confidential instances, and 3) a self-supervised pretext component to aid the learning process. We have systematically evaluated our method on the CT images of 229 COVID-19 cases, including 50 severe and 179 non-severe cases. Our method could obtain an average accuracy of 95.8%, with 93.6% sensitivity and 96.4% specificity, which outperformed previous works.

eess.IV↗

arXiv2021

Sharp Hardy inequalities via Riemannian submanifolds

This paper is devoted to Hardy inequalities concerning distance functions from submanifolds of arbitrary codimensions in the Riemannian setting. On a Riemannian manifold with non-negative curvature, we establish several sharp weighted Hardy inequalities in the cases when the submanifold is compact as well as non-compact. In particular, these inequalities remain valid even if the ambient manifold is compact, in which case we find an optimal space of smooth functions to study Hardy inequalities. Further examples are also provided. Our results complement in several aspects those obtained recently in the Euclidean and Riemannian settings.

math.DG↗

arXiv2021

A Two-Stage Wavelet Decomposition Method for Instantaneous Power Quality Indices Estimation Considering Interharmonics and Transient Disturbances

As the complexity increases in modern power systems, power quality analysis considering interharmonics has become a challenging and important task. This paper proposes a novel decomposition and estimation method for instantaneous power quality indices (PQIs) monitoring in single-phase and three-phase systems with interharmonics and transient disturbances. To separate the interharmonic components, a set of new scaling filter and wavelet filter with narrow transition bands are designed for the undecimated wavelet packet transform (UWPT). Further, a two-stage decomposition method for multi-tone voltage and current signals is proposed. The Hilbert transform (HT) is applied to calculate the instantaneous amplitude and phase of each frequency component, which accordingly allows the monitoring of different PQI parameters. Numerical tests are conducted to check the performance of the proposed method. The test results show that compared to other conventional approaches, instantaneous PQIs estimated by the proposed method present significant advances for tracking transitory changes in power systems, and could be considered as a helpful tool for high-accuracy PQ detections.

eess.SP↗

arXiv2021

Resolution of the paradox of the diamagnetic effect on the Kibble coil

Employing very simple electro-mechanical principles known from classical physics, the Kibble balance establishes a very precise and absolute link between quantum electrical standards and macroscopic mass or force measurements. The success of the Kibble balance, in both determining fundamental constants ($h$, $N_A$, $e$) and realizing a quasi-quantum mass in the 2019 newly revised International System of Units, relies on the perfection of Maxwell's equations and the symmetry they describe between Lorentz's force and Faraday's induction, a principle and a symmetry stunningly demonstrated in the weighing and velocity modes of Kibble balances to within $1\times10^{-8}$, with nothing but imperfect wires and magnets. However, recent advances in the understanding of the current effect in Kibble balances reveal a troubling paradox. A diamagnetic effect, a force that does not cancel between mass-on and mass-off measurement, is challenging balance maker's assumptions of symmetry at levels that are almost two orders of magnitude larger than the reported uncertainties. The diamagnetic effect, if it exists, shows up in weighing mode without a readily apparent reciprocal effect in the velocity mode, begging questions about systematic errors at the very foundation of the new measurement system. The hypothetical force is caused by the coil current changing the magnetic field, producing an unaccounted force that is systematically modulated with the weighing current. Here we show that this diamagnetic force exists, but the additional force does not change the equivalence between weighing and velocity measurements. We reveal the unexpected way that symmetry is preserved and show that for typical materials and geometries the total relative effect on the measurement is $\approx 1\times10^{-9}$.

physics.ins-det↗

arXiv2020

Regularized Attentive Capsule Network for Overlapped Relation Extraction

Distantly supervised relation extraction has been widely applied in knowledge base construction due to its less requirement of human efforts. However, the automatically established training datasets in distant supervision contain low-quality instances with noisy words and overlapped relations, introducing great challenges to the accurate extraction of relations. To address this problem, we propose a novel Regularized Attentive Capsule Network (RA-CapNet) to better identify highly overlapped relations in each informal sentence. To discover multiple relation features in an instance, we embed multi-head attention into the capsule network as the low-level capsules, where the subtraction of two entities acts as a new form of relation query to select salient features regardless of their positions. To further discriminate overlapped relation features, we devise disagreement regularization to explicitly encourage the diversity among both multiple attention heads and low-level capsules. Extensive experiments conducted on widely used datasets show that our model achieves significant improvements in relation extraction.

cs.CL↗

arXiv2020

Compressive Circular Polarization Snapshot Spectral Imaging

A compressive sensing based circular polarization snapshot spectral imaging system is proposed in this paper to acquire two-dimensional spatial, one-dimensional circular polarization (the right and left circular polarization), and one-dimensional spectral information, simultaneously. Using snapshot can collect the entire four-dimensional datacube in a single integration period. The dispersion prism in the coded aperture snapshot spectral imager is replaced by the combination of an Amici prism and a Wollaston prism to implement the spectral shifting along two orthogonal directions, which greatly improves the spectral resolution of the image. The right and left circular polarization components of objects are extracted by the assemble with an achromatic quarter wave-plate and a Wollaston prism. The encoding and reconstruction are illustrated comprehensively. The feasibility is verified by the simulation. It provides us an alternative approach for circular polarization spectral imaging such as defogging, underwater imaging, and so on.

physics.optics↗

arXiv2020

SE-ECGNet: A Multi-scale Deep Residual Network with Squeeze-and-Excitation Module for ECG Signal Classification

The classification of electrocardiogram (ECG) signals, which takes much time and suffers from a high rate of misjudgment, is recognized as an extremely challenging task for cardiologists. The major difficulty of the ECG signals classification is caused by the long-term sequence dependencies. Most existing approaches for ECG signal classification use Recurrent Neural Network models, e.g., LSTM and GRU, which are unable to extract accurate features for such long sequences. Other approaches utilize 1-Dimensional Convolutional Neural Network (CNN), such as ResNet or its variant, and they can not make good use of the multi-lead information from ECG signals.Based on the above observations, we develop a multi-scale deep residual network for the ECG signal classification task. We are the first to propose to treat the multi-lead signal as a 2-dimensional matrix and combines multi-scale 2-D convolution blocks with 1-D convolution blocks for feature extraction. Our proposed model achieves 99.2% F1-score in the MIT-BIH dataset and 89.4% F1-score in Alibaba dataset and outperforms the state-of-the-art performance by 2% and 3%, respectively, view related code and data at https://github.com/Amadeuszhao/SE-ECGNet

cs.LG↗

arXiv2020

Noise2Context: Context-assisted Learning 3D Thin-layer Low Dose CT Without Clean Data

Computed tomography (CT) has played a vital role in medical diagnosis, assessment, and therapy planning, etc. In clinical practice, concerns about the increase of X-ray radiation exposure attract more and more attention. To lower the X-ray radiation, low-dose CT is often used in certain scenarios, while it will induce the degradation of CT image quality. In this paper, we proposed a training method that trained denoising neural networks without any paired clean data. we trained the denoising neural network to map one noise LDCT image to its two adjacent LDCT images in a singe 3D thin-layer low-dose CT scanning, simultaneously In other words, with some latent assumptions, we proposed an unsupervised loss function with the integration of the similarity between adjacent CT slices in 3D thin-layer lowdose CT to train the denoising neural network in an unsupervised manner. For 3D thin-slice CT scanning, the proposed virtual supervised loss function was equivalent to a supervised loss function with paired noisy and clean samples when the noise in the different slices from a single scan was uncorrelated and zero-mean. Further experiments on Mayo LDCT dataset and a realistic pig head were carried out and demonstrated superior performance over existing unsupervised methods.

eess.IV↗