arXiv subjects

Hang Zhang

Publications and source records attributed to Hang Zhang.

At least 217 records · Page 12Linked to original sources

arXiv2020

Bayesian Learning of Probabilistic Dipole Inversion for Quantitative Susceptibility Mapping

A learning-based posterior distribution estimation method, Probabilistic Dipole Inversion (PDI), is proposed to solve quantitative susceptibility mapping (QSM) inverse problem in MRI with uncertainty estimation. A deep convolutional neural network (CNN) is used to represent the multivariate Gaussian distribution as the approximated posterior distribution of susceptibility given the input measured field. In PDI, such CNN is firstly trained on healthy subjects dataset with labels by maximizing the posterior Gaussian distribution loss function as used in Bayesian deep learning. When tested on new dataset without any label, PDI updates the pre-trained network in an unsupervised fashion by minimizing the KL divergence between the approximated posterior distribution represented by CNN and the true posterior distribution given the likelihood distribution from known physical model and prior distribution. Based on our experiments, PDI provides additional uncertainty estimation compared to the conventional MAP approach, meanwhile addressing the potential discrepancy issue of CNN when test data deviates from training dataset.

eess.IV↗

arXiv2020

An Efficient Index Method for the Optimal Route Query over Multi-Cost Networks

Smart city has been consider the wave of the future and the route recommendation in networks is a fundamental problem in it. Most existing approaches for the shortest route problem consider that there is only one kind of cost in networks. However, there always are several kinds of cost in networks and users prefer to select an optimal route under the global consideration of these kinds of cost. In this paper, we study the problem of finding the optimal route in the multi-cost networks. We prove this problem is NP-hard and the existing index techniques cannot be used to this problem. We propose a novel partition-based index with contour skyline techniques to find the optimal route. We propose a vertex-filtering algorithm to facilitate the query processing. We conduct extensive experiments on six real-life networks and the experimental results show that our method has an improvement in efficiency by an order of magnitude compared to the previous heuristic algorithms.

cs.DB↗

arXiv2020

FairNN- Conjoint Learning of Fair Representations for Fair Decisions

In this paper, we propose FairNN a neural network that performs joint feature representation and classification for fairness-aware learning. Our approach optimizes a multi-objective loss function in which (a) learns a fair representation by suppressing protected attributes (b) maintains the information content by minimizing a reconstruction loss and (c) allows for solving a classification task in a fair manner by minimizing the classification error and respecting the equalized odds-based fairness regularized. Our experiments on a variety of datasets demonstrate that such a joint approach is superior to separate treatment of unfairness in representation learning or supervised learning. Additionally, our regularizers can be adaptively weighted to balance the different components of the loss function, thus allowing for a very general framework for conjoint fair representation learning and decision making.

cs.LG↗

arXiv2020

AutoGluon-Tabular: Robust and Accurate AutoML for Structured Data

We introduce AutoGluon-Tabular, an open-source AutoML framework that requires only a single line of Python to train highly accurate machine learning models on an unprocessed tabular dataset such as a CSV file. Unlike existing AutoML frameworks that primarily focus on model/hyperparameter selection, AutoGluon-Tabular succeeds by ensembling multiple models and stacking them in multiple layers. Experiments reveal that our multi-layer combination of many models offers better use of allocated training time than seeking out the best. A second contribution is an extensive evaluation of public and commercial AutoML platforms including TPOT, H2O, AutoWEKA, auto-sklearn, AutoGluon, and Google AutoML Tables. Tests on a suite of 50 classification and regression tasks from Kaggle and the OpenML AutoML Benchmark reveal that AutoGluon is faster, more robust, and much more accurate. We find that AutoGluon often even outperforms the best-in-hindsight combination of all of its competitors. In two popular Kaggle competitions, AutoGluon beat 99% of the participating data scientists after merely 4h of training on the raw data.

stat.ML↗

arXiv2020

RSANet: Recurrent Slice-wise Attention Network for Multiple Sclerosis Lesion Segmentation

Brain lesion volume measured on T2 weighted MRI images is a clinically important disease marker in multiple sclerosis (MS). Manual delineation of MS lesions is a time-consuming and highly operator-dependent task, which is influenced by lesion size, shape and conspicuity. Recently, automated lesion segmentation algorithms based on deep neural networks have been developed with promising results. In this paper, we propose a novel recurrent slice-wise attention network (RSANet), which models 3D MRI images as sequences of slices and captures long-range dependencies through a recurrent manner to utilize contextual information of MS lesions. Experiments on a dataset with 43 patients show that the proposed method outperforms the state-of-the-art approaches. Our implementation is available online at https://github.com/tinymilky/RSANet.

eess.IV↗

arXiv2020

GluonCV and GluonNLP: Deep Learning in Computer Vision and Natural Language Processing

We present GluonCV and GluonNLP, the deep learning toolkits for computer vision and natural language processing based on Apache MXNet (incubating). These toolkits provide state-of-the-art pre-trained models, training scripts, and training logs, to facilitate rapid prototyping and promote reproducible research. We also provide modular APIs with flexible building blocks to enable efficient customization. Leveraging the MXNet ecosystem, the deep learning models in GluonCV and GluonNLP can be deployed onto a variety of platforms with different programming languages. The Apache 2.0 license has been adopted by GluonCV and GluonNLP to allow for software distribution, modification, and usage.

cs.LG↗

arXiv2019

Slow light,double directional rainbow trapping and releasing at terahertz frequencies

Slow light and rainbow trapping attract many attentions in last twenty years, and in most of the previous works, the researchers achieve the slow light and rainbow trapping with complicate configurations or techniques, for example, metamaterial techniques. In this paper, we propose a simple waveguide consisted with semiconductor and dielectric layers, and we find that the slow-light peaks appear in the dispersion curves. Besides, the cutoff frequency with $v_g = 0$ ($v_g$ is the group velocity) in the slow-light peaks is depending on the thicknesses of the semiconductor and dielectric. Then, we design a tapered, horizontally symmetric structure. By using software COMSOL and finite difference time domain method as well, we achieve double directional rainbow trapping and more importantly, releasing the rainbow in such structure. This approach is promising for optical isolator, optical buffer, optical switch and other optical functional devices in miniaturization optical integrated circuit.

physics.optics↗

arXiv2019

Ultra-subwavelength focusing and giant magnetic-field enhancement in a low-loss one-way waveguide based on remanence

The subwavelength focusing based on surface plasmon polaritons (SPP) has been widely explored in tapered metallic structures. However, the efficiency of energy localization is relatively weak, largely due to high propagation loss and strong back reflection. Here, we propose a straight-tapered 3-dimensional (3D) one-way surface magnetoplasmon (SMP) waveguide with the ending surface of $\sim 10^{-4}λ_0 \times 10^{-4}λ_0$ to achieve energy focusing in the microwave regime. Due to low propagation loss of SMP, we achieve huge magnetic field enhancement in such an ultra-subwavelength area, by five orders of magnitude. Instead of using an external static magnetic field, our proposed SMP waveguide relies on remanence, which is very convenient for operating practical 3D applications. These results show promising applications in magnetic-field enhancing or quenching fluorescence, luminescence or nonlinearity of 2D materials, novel scanning near-field microwave microscopy and energy storage.

physics.optics↗

arXiv2019

2nd Place Solution to the GQA Challenge 2019

We present a simple method that achieves unexpectedly superior performance for Complex Reasoning involved Visual Question Answering. Our solution collects statistical features from high-frequency words of all the questions asked about an image and use them as accurate knowledge for answering further questions of the same image. We are fully aware that this setting is not ubiquitously applicable, and in a more common setting one should assume the questions are asked separately and they cannot be gathered to obtain a knowledge base. Nonetheless, we use this method as an evidence to demonstrate our observation that the bottleneck effect is more severe on the feature extraction part than it is on the knowledge reasoning part. We show significant gaps when using the same reasoning model with 1) ground-truth features; 2) statistical features; 3) detected features from completely learned detectors, and analyze what these gaps mean to researches on visual reasoning topics. Our model with the statistical features achieves the 2nd place in the GQA Challenge 2019.

cs.CV↗

arXiv2019

Compressive Sensing with a Multiple Convex Sets Domain

In this paper, we study a general framework for compressive sensing assuming the existence of the prior knowledge that $\boldsymbol{x}^*$ belongs to the union of multiple convex sets, $\boldsymbol{x}^{*} \in \bigcup_i \mathcal{C}_i$. In fact, by proper choices of these convex sets in the above framework, the problem can be transformed to well known CS problems such as the phase retrieval, quantized compressive sensing, and model-based CS. First we analyze the impact of this prior knowledge on the minimum number of measurements $M$ to guarantee the uniqueness of the solution. Then we formulate a universal objective function for signal recovery, which is both computationally inexpensive and flexible. Then, an algorithm based on \emph{multiplicative weight update} and \emph{proximal gradient descent} is proposed and analyzed for computation and its properties are analyzed for signal reconstruction. Finally, we investigate as to how we can improve the signal recovery by introducing regularizers into the objective function.

math.OC↗

arXiv2019

A Computer Vision Application for Assessing Facial Acne Severity from Selfie Images

We worked with Nestle SHIELD (Skin Health, Innovation, Education, and Longevity Development, NSH) to develop a deep learning model that is able to assess acne severity from selfie images as accurate as dermatologists. The model was deployed as a mobile application, providing patients an easy way to assess and track the progress of their acne treatment. NSH acquired 4,700 selfie images for this study and recruited 11 internal dermatologists to label them in five categories: 1-Clear, 2- Almost Clear, 3-Mild, 4-Moderate, 5-Severe. Using OpenCV to detect facial landmarks we cut specific skin patches from the selfie images in order to minimize irrelevant background. We then applied a transfer learning approach by extracting features from the patches using a ResNet 152 pre-trained model, followed by a fully connected layer trained to approximate the desired severity rating. To address the problem of spatial sensitivity of CNN models, we introduce a new image rolling data augmentation approach, effectively causing acne lesions appeared in more locations in the training images. Our results demonstrate that this approach improved the generalization of the CNN model, outperforming more than half of the panel of human dermatologists on test images. To our knowledge, this is the first deep learning-based solution for acne assessment using selfie images.

cs.CV↗

arXiv2019

A Light-Driven Microgel Rotor

The current understanding of motility through body shape deformation of microorganisms and the knowledge of fluid flows at the microscale provides ample examples for mimicry and design of soft microrobots. In this work, a two-dimensional spiral is presented that is capable of rotating by non-reciprocal curling deformations. The body of the microswimmer is a ribbon consisting of a thermo-responsive hydrogel bilayer with embedded plasmonic gold nanorods. Such a system allows fast local photothermal heating and non-reciprocal bending deformation of the hydrogel bilayer under non-equilibrium conditions. We show that the spiral acts as a spring capable of large deformations thanks to its low stiffness, which is tunable by the swelling degree of the hydrogel and the temperature. Tethering the ribbon to a freely rotating microsphere enables rotational motion of the spiral by stroboscopic irradiation. The efficiency of the rotor is estimated using resistive force theory for Stokes flow. The present research demonstrates microscopic locomotion by the shape change of a spiral and may find applications in the field of microfluidics, or soft micro-robotics.

cond-mat.mtrl-sci↗

arXiv2019

Fidelity Imposed Network Edit (FINE) for Solving Ill-Posed Image Reconstruction

Deep learning (DL) is increasingly used to solve ill-posed inverse problems in imaging, such as reconstruction from noisy or incomplete data, as DL offers advantages over explicit image feature extractions in defining the needed prior. However, DL typically does not incorporate the precise physics of data generation or data fidelity. Instead, DL networks are trained to output some average response to an input. Consequently, DL image reconstruction contains errors, and may perform poorly when the test data deviates significantly from the training data, such as having new pathological features. To address this lack of data fidelity problem in DL image reconstruction, a novel approach, which we call fidelity-imposed network edit (FINE), is proposed. In FINE, a pre-trained prior network's weights are modified according to the physical model, on a test case. Our experiments demonstrate that FINE can achieve superior performance in two important inverse problems in neuroimaging: quantitative susceptibility mapping (QSM) and under-sampled reconstruction in MRI.

eess.IV↗

arXiv2019

Dynamic Mini-batch SGD for Elastic Distributed Training: Learning in the Limbo of Resources

With an increasing demand for training powers for deep learning algorithms and the rapid growth of computation resources in data centers, it is desirable to dynamically schedule different distributed deep learning tasks to maximize resource utilization and reduce cost. In this process, different tasks may receive varying numbers of machines at different time, a setting we call elastic distributed training. Despite the recent successes in large mini-batch distributed training, these methods are rarely tested in elastic distributed training environments and suffer degraded performance in our experiments, when we adjust the learning rate linearly immediately with respect to the batch size. One difficulty we observe is that the noise in the stochastic momentum estimation is accumulated over time and will have delayed effects when the batch size changes. We therefore propose to smoothly adjust the learning rate over time to alleviate the influence of the noisy momentum estimation. Our experiments on image classification, object detection and semantic segmentation have demonstrated that our proposed Dynamic SGD method achieves stabilized performance when varying the number of GPUs from 8 to 128. We also provide theoretical understanding on the optimality of linear learning rate scheduling and the effects of stochastic momentum.

cs.LG↗

arXiv2019

Bag of Freebies for Training Object Detection Neural Networks

Training heuristics greatly improve various image classification model accuracies~\cite{he2018bag}. Object detection models, however, have more complex neural network structures and optimization targets. The training strategies and pipelines dramatically vary among different models. In this works, we explore training tweaks that apply to various models including Faster R-CNN and YOLOv3. These tweaks do not change the model architectures, therefore, the inference costs remain the same. Our empirical results demonstrate that, however, these freebies can improve up to 5% absolute precision compared to state-of-the-art baselines.

cs.CV↗

arXiv2019

Blind Estimation Algorithms for I/Q Imbalance in Direct Down-conversion Receivers

As known, receivers with in-phase and quadrature phase (I/Q) down conversion, especially direct-conversion architectures, always suffer from I/Q imbalance. I/Q imbalance is caused by amplitude and phase mismatch between I/Q paths. The performance degradation resulting from I/Q imbalance can not be mitigated with simply higher signal to noise ratio (SNR). Thus, I/Q imbalance compensation in the digital domain is critical. There are two main contributions in this paper. Firstly, we proposed a blind estimation algorithm for I/Q imbalance parameters based on joint first and second order statistics (FSS) which has lower complexity than conventional Gaussian maximum likelihood estimation (GMLE). This can be used for further processing such as equalization in the presence of receiver IQ imbalance. In addition, we find out the reason of the error floor in conventional I/Q imbalance compensation method based on the conjugate signal model (CSM). The proposed joint first order statistics and conjugate signal model (FSCSM) compensation algorithm can reach the ideal bit error rate (BER) performance.

eess.SP↗

arXiv2019

Workflow Scheduling in the Cloud with Weighted Upward-rank Priority Scheme Using Random Walk and Uniform Spare Budget Splitting

We study a difficult problem of how to schedule complex workflows with precedence constraints under a limited budget in the cloud environment. We first formulate the scheduling problem as an integer programming problem, which can be optimized and used as the baseline of performance. We then consider the traditional approach of scheduling jobs in a prioritized order based on the upward-rank of each job. For those jobs with no precedence constraints among themselves, the plain upward-rank priority scheme assigns priorities in an arbitrary way. We propose a job prioritization scheme that uses Markovian chain stationary probabilities as a measure of importance of jobs. The scheme keeps the precedence order for the jobs that have precedence constraints between each other, and assigns priorities according to the jobs' importance for the jobs without precedence constraints. We finally design a uniform spare budget splitting strategy, which splits the spare budget uniformly across all the jobs. We test our algorithms on a variety of workflows, including FFT, Gaussian elimination, typical scientific workflows, randomly generated workflows and workflows from an in-production cluster of an online streaming service company. We compare our algorithms with the-state-of-art algorithms. The empirical results show that the uniform spare budget splitting scheme outperforms the splitting scheme in proportion to extra demand in average for most cases, and the Markovian based prioritization further improves the workflow makespan.

cs.DC↗

arXiv2019

Blaze: Simplified High Performance Cluster Computing

MapReduce and its variants have significantly simplified and accelerated the process of developing parallel programs. However, most MapReduce implementations focus on data-intensive tasks while many real-world tasks are compute intensive and their data can fit distributedly into the memory. For these tasks, the speed of MapReduce programs can be much slower than those hand-optimized ones. We present Blaze, a C++ library that makes it easy to develop high performance parallel programs for such compute intensive tasks. At the core of Blaze is a highly-optimized in-memory MapReduce function, which has three main improvements over conventional MapReduce implementations: eager reduction, fast serialization, and special treatment for a small fixed key range. We also offer additional conveniences that make developing parallel programs similar to developing serial programs. These improvements make Blaze an easy-to-use cluster computing library that approaches the speed of hand-optimized parallel code. We apply Blaze to some common data mining tasks, including word frequency count, PageRank, k-means, expectation maximization (Gaussian mixture model), and k-nearest neighbors. Blaze outperforms Apache Spark by more than 10 times on average for these tasks, and the speed of Blaze scales almost linearly with the number of nodes. In addition, Blaze uses only the MapReduce function and 3 utility functions in its implementation while Spark uses almost 30 different parallel primitives in its official implementation.

cs.DC↗