arXiv subjects

Lin Yang

Publications and source records attributed to Lin Yang.

At least 271 records · Page 15Linked to original sources

arXiv2017

MDNet: A Semantically and Visually Interpretable Medical Image Diagnosis Network

The inability to interpret the model prediction in semantically and visually meaningful ways is a well-known shortcoming of most existing computer-aided diagnosis methods. In this paper, we propose MDNet to establish a direct multimodal mapping between medical images and diagnostic reports that can read images, generate diagnostic reports, retrieve images by symptom descriptions, and visualize attention, to provide justifications of the network diagnosis process. MDNet includes an image model and a language model. The image model is proposed to enhance multi-scale feature ensembles and utilization efficiency. The language model, integrated with our improved attention mechanism, aims to read and explore discriminative image feature descriptions from reports to learn a direct mapping from sentence words to image pixels. The overall network is trained end-to-end by using our developed optimization strategy. Based on a pathology bladder cancer images and its diagnostic reports (BCIDR) dataset, we conduct sufficient experiments to demonstrate that MDNet outperforms comparative baselines. The proposed image model obtains state-of-the-art performance on two CIFAR datasets as well.

cs.CV↗

arXiv2017

Suggestive Annotation: A Deep Active Learning Framework for Biomedical Image Segmentation

Image segmentation is a fundamental problem in biomedical image analysis. Recent advances in deep learning have achieved promising results on many biomedical image segmentation benchmarks. However, due to large variations in biomedical images (different modalities, image settings, objects, noise, etc), to utilize deep learning on a new application, it usually needs a new set of training data. This can incur a great deal of annotation effort and cost, because only biomedical experts can annotate effectively, and often there are too many instances in images (e.g., cells) to annotate. In this paper, we aim to address the following question: With limited effort (e.g., time) for annotation, what instances should be annotated in order to attain the best performance? We present a deep active learning framework that combines fully convolutional network (FCN) and active learning to significantly reduce annotation effort by making judicious suggestions on the most effective annotation areas. We utilize uncertainty and similarity information provided by FCN and formulate a generalized version of the maximum set cover problem to determine the most representative and uncertain areas for annotation. Extensive experiments using the 2015 MICCAI Gland Challenge dataset and a lymph node ultrasound image segmentation dataset show that, using annotation suggestions by our method, state-of-the-art segmentation performance can be achieved by using only 50% of training data.

cs.CV↗

arXiv2017

Cycloidal magnetism driven ferroelectricity in double tungstate LiFe(WO$_4$)$_2$

Tungstates $A$WO$_4$ with the wolframite structure characterized by the $A$O$_6$ octahedral zigzag chains along the $c$-axis, can be magnetic if $A$=Mn, Fe, Co, Cu, Ni. Among them, MnWO$_4$ is a unique member with a cycloid Mn$^{2+}$ spin order developed at low temperature, leading to an interesting type-II multiferroic behavior. However, so far no other multiferroic material in the tungstate family has been found. In this work, we present the synthesis and the systematic study of the double tungstate LiFe(WO$_4$)$_2$. Experimental characterizations including structural, thermodynamic, magnetic, neutron powder diffraction, and pyroelectric measurements, unambiguously confirm that LiFe(WO$_4$)$_2$ is the secondly found multiferroic system in the tungstate family. The cycloidal magnetism driven ferroelectricity is also verified by density functional theory calculations. Although here the magnetic couplings between Fe ions are indirect, namely via the so-called super-super-exchanges, the temperatures of magnetic and ferroelectric transitions are surprisingly much higher than those of MnWO$_4$.

cond-mat.mtrl-sci↗

arXiv2017

Max-Margin based Discriminative Feature Learning

In this paper, we propose a new max-margin based discriminative feature learning method. Specifically, we aim at learning a low-dimensional feature representation, so as to maximize the global margin of the data and make the samples from the same class as close as possible. In order to enhance the robustness to noise, a $l_{2,1}$ norm constraint is introduced to make the transformation matrix in group sparsity. In addition, for multi-class classification tasks, we further intend to learn and leverage the correlation relationships among multiple class tasks for assisting in learning discriminative features. The experimental results demonstrate the power of the proposed method against the related state-of-the-art methods.

cs.LG↗

arXiv2017

Increasing Trends of Guillain-Barré Syndrome (GBS) and Dengue in Hong Kong

Background: Guillain-Barré Syndrome (GBS) is a common type of severe acute paralytic neuropathy and associated with other virus infections such as dengue fever and Zika. This study investigate the relationship between GBS, dengue, local meteorological factors in Hong Kong and global climatic factors from January 2000 to June 2016. Methods: The correlations between GBS, dengue, Multivariate El Nino Southern Oscillation Index (MEI) and local meteorological data were explored by the Spearman Rank correlations and cross-correlations between these time series. Poisson regression models were fitted to identify nonlinear associations between MEI and dengue. Cross wavelet analysis was applied to infer potential non-stationary oscillating associations among MEI, dengue and GBS. Findings : An increasing trend was found for both GBS cases and imported dengue cases in Hong Kong. We found a weak but statistically significant negative correlation between GBS and local meteorological factors. MEI explained over 12\% of dengue's variations from Poisson regression models. Wavelet analyses showed that there is possible non-stationary oscillating association between dengue and GBS from 2005 to 2015 in Hong Kong. Our study has led to an improved understanding of the timing and relationship between GBS, dengue and MEI.

q-bio.PE↗

arXiv2016

Microscopic Muscle Image Enhancement

We propose a robust image enhancement algorithm dedicated for muscle fiber specimen images captured by optical microscopes. Blur or out of focus problems are prevalent in muscle images during the image acquisition stage. Traditional image deconvolution methods do not work since they assume the blur kernels are known and also produce ring artifacts. We provide a compact framework which involves a novel spatially non-uniform blind deblurring approach specialized to muscle images which automatically detects and alleviates degraded regions. Ring artifacts problems are addressed and a kernel propagation strategy is proposed to speedup the algorithm and deals with the high non-uniformity of the blur kernels on muscle images. Experiments show that the proposed framework performs well on muscle images taken with modern advanced optical microscopes. Our framework is free of laborious parameter settings and is computationally efficient.

cs.CV↗

arXiv2016

Online Offering Strategies for Storage-Assisted Renewable Power Producer in Hour-Ahead Market

A promising approach to hedge against the inherent uncertainty of renewable generation is to equip the renewable plants with energy storage systems. This paper focuses on designing profit maximization offering strategies, i.e., the strategies that determine the offering price and volume, for a storage-assisted renewable power producer that participates in hour-ahead electricity market. Designing the strategies is challenging since (i) the underlying problem is coupled across time due to the evolution of the storage level, and (ii) inputs to the problem including the renewable output and market clearing price are unknown when submitting offers. Following the competitive online algorithm design approach, we first study a basic setting where the renewable output and the clearing price are known for the next hour. We propose sOffer, a simple online offering strategy that achieves the best possible competitive ratio of O(log θ), where $θ$ is the ratio between the maximum and the minimum clearing prices. Then, we consider the case where the clearing price is unknown. By exploiting the idea of submitting multiple offers to combat price uncertainty, we propose mOffer, and demonstrate that the competitive ratio of mOffer converges to that of sOffer as the number of offers grows. Finally, we extend our approach to the scenario where the renewable output has forecasting error. We propose gOffer as the generalized offering strategy and characterize its competitive ratio as a function of the forecasting error. Our trace-driven experiments demonstrate that our algorithms achieve performance close to the offline optimal and outperform a baseline alternative significantly.

cs.GT↗

arXiv2016

Forward Backward Similarity Search in Knowledge Networks

Similarity search is a fundamental problem in social and knowledge networks like GitHub, DBLP, Wikipedia, etc. Existing network similarity measures are limited because they only consider similarity from the perspective of the query node. However, due to the complicated topology of real-world networks, ignoring the preferences of target nodes often results in odd or unintuitive performance. In this work, we propose a dual perspective similarity metric called Forward Backward Similarity (FBS) that efficiently computes topological similarity from the perspective of both the query node and the perspective of candidate nodes. The effectiveness of our method is evaluated by traditional quantitative ranking metrics and large-scale human judgement on four large real world networks. The proposed method matches human preference and outperforms other similarity search algorithms on community overlap and link prediction. Finally, we demonstrate top-5 rankings for five famous researchers on an academic collaboration network to illustrate how our approach captures semantics more intuitively than other approaches.

cs.SI↗

arXiv2016

Decreased aneurysmal subarachnoid hemorrhage incidence rate in elderly population than in middle aged population: a retrospective analysis of 8,144 cases in Mainland China

Purpose: Rupture of an intracranial aneurysm is the most common cause of subarachnoid haemorrhage (SAH), which is a life-threatening acute cerebrovascular event that typically affects working-age people. This study aims to investigate the aneurysmal SAH incidence rate in elderly population than in middle aged population in China. Materials and methods: Aneurysmal SAH cases were collected retrospectively from the archives of 21 hospitals in Mainland China. All the cases collected were from September 2016 and backward consecutively for a period of time up to 8 years. SAH was initially diagnosed by brain computed tomography, and CT angiography (CTA) or digital subtraction angiography (DSA) was followed and SAH was confirmed to be due to cerebral aneurysm. When for cases multiple bleeding occurred, the age of the first SAH was used in this study. The toltal incidence from all hospital at each age were summed together for females and males; then adjusted by the total population number at each age for females and males. The total population data was from the 2010 population census of the People's Republic of China. Results: In total there were 8,144 cases, with 4,861 females and 3,283 males. Our analysis shows for both females and males the relative aneurysmal SAH rate started to decrease after around 65 years old. The males the relative aneurysmal SAH rate might have started to decrease after around 55 years old. Conclusion: In contrast to previous reports, our data demonstrated a decreased aneurysmal subarachnoid hemorrhage incidence rate in elderly population than in middle aged population. Our data therefore support the hypothesis that aneurysms do not grow progressively once they form but probably either rupture or stabilize and that very elderly patients are at a reduced risk of rupture compared with atients who are younger with the same-sized aneurysms.

q-bio.TO↗

arXiv2016

Combining Fully Convolutional and Recurrent Neural Networks for 3D Biomedical Image Segmentation

Segmentation of 3D images is a fundamental problem in biomedical image analysis. Deep learning (DL) approaches have achieved state-of-the-art segmentation perfor- mance. To exploit the 3D contexts using neural networks, known DL segmentation methods, including 3D convolution, 2D convolution on planes orthogonal to 2D image slices, and LSTM in multiple directions, all suffer incompatibility with the highly anisotropic dimensions in common 3D biomedical images. In this paper, we propose a new DL framework for 3D image segmentation, based on a com- bination of a fully convolutional network (FCN) and a recurrent neural network (RNN), which are responsible for exploiting the intra-slice and inter-slice contexts, respectively. To our best knowledge, this is the first DL framework for 3D image segmentation that explicitly leverages 3D image anisotropism. Evaluating using a dataset from the ISBI Neuronal Structure Segmentation Challenge and in-house image stacks for 3D fungus segmentation, our approach achieves promising results comparing to the known DL-based 3D segmentation approaches.

cs.CV↗

arXiv2016

SemiContour: A Semi-supervised Learning Approach for Contour Detection

Supervised contour detection methods usually require many labeled training images to obtain satisfactory performance. However, a large set of annotated data might be unavailable or extremely labor intensive. In this paper, we investigate the usage of semi-supervised learning (SSL) to obtain competitive detection accuracy with very limited training data (three labeled images). Specifically, we propose a semi-supervised structured ensemble learning approach for contour detection built on structured random forests (SRF). To allow SRF to be applicable to unlabeled data, we present an effective sparse representation approach to capture inherent structure in image patches by finding a compact and discriminative low-dimensional subspace representation in an unsupervised manner, enabling the incorporation of abundant unlabeled patches with their estimated structured labels to help SRF perform better node splitting. We re-examine the role of sparsity and propose a novel and fast sparse coding algorithm to boost the overall learning efficiency. To the best of our knowledge, this is the first attempt to apply SSL for contour detection. Extensive experiments on the BSDS500 segmentation dataset and the NYU Depth dataset demonstrate the superiority of the proposed method.

cs.CV↗

arXiv2015

On Hypothesis Testing for Poisson Processes. Regular Case

We consider the problem of hypothesis testing in the situation when the first hypothesis is simple and the second one is local one-sided composite. We describe the choice of the thresholds and the power functions of the Score Function test, of the General Likelihood Ratio test, of the Wald test and of two Bayes tests in the situation when the intensity function of the observed inhomogeneous Poisson process is smooth with respect to the parameter. It is shown that almost all these tests are asymptotically uniformly most powerful. The results of numerical simulations are presented.

math.ST↗

arXiv2015

On a Poissonian Change-Point Model with Variable Jump Size

A model of Poissonian observation having a jump (change-point) in the intensity function is considered. Two cases are studied. The first one corresponds to the situation when the jump size converges to a non-zero limit, while in the second one the limit is zero. The limiting likelihood ratios in these two cases are quite different. In the first case, like in the case of a fixed jump size, the normalized likelihood ratio converges to a log Poisson process. In the second case, the normalized likelihood ratio converges to a log Wiener process, and so, the statistical problems of parameter estimation and hypotheses testing are asymptotically equivalent in this case to the well known problems of change-point estimation and testing for the model of a signal in white Gaussian noise. The properties of the maximum likelihood and Bayesian estimators, as well as those of the general likelihood ratio, Wald's and Bayesian tests are deduced form the convergence of normalized likelihood ratios. The convergence of the moments of the estimators is also established. The obtained theoretical results are illustrated by numerical simulations.

math.ST↗

arXiv2015

On Hypothesis Testing for Poisson Processes. Singular Cases

We consider the problem of hypothesis testing in the situation where the first hypothesis is simple and the second one is local one-sided composite. We describe the choice of the thresholds and the power functions of different tests when the intensity function of the observed inhomogeneous Poisson process has two different types of singularity: cusp and discontinuity. The asymptotic results are illustrated by numerical simulations.

math.ST↗

arXiv2014

Global Spatio-temporal Patterns of Influenza in the Post-pandemic Era

We study the global spatio-temporal patterns of influenza dynamics. This is achieved by analysing and modelling weekly laboratory confirmed cases of influenza A and B from 138 countries between January 2006 and May 2014. The data were obtained from FluNet, the surveillance network compiled by the the World Health Organization. We report a pattern of {\it skip-and-resurgence} behavior between the years 2011 and 2013 for influenza H1N1/09, the strain responsible for the 2009 pandemic, in Europe and Eastern Asia. In particular, the expected H1N1/09 epidemic outbreak in 2011 failed to occur (or"skipped") in many countries across the globe, although an outbreak occurred in the following year. We also report a pattern of {\it well-synchronized} 2010 winter wave of H1N1/09 in the Northern Hemisphere countries, and a pattern of replacement of strain H1N1/77 by H1N1/09 between the 2009 and 2012 influenza seasons. Using both a statistical and a mechanistic mathematical model, and through fitting the data of 108 countries (108 countries in a statistical model and 10 large populations with a mechanistic model), we discuss the mechanisms that are likely to generate these events taking into account the role of multi-strain dynamics. A basic understanding of these patterns has important public health implications and scientific significance.

q-bio.PE↗

arXiv2014

Dimensional evolution between one- and two-dimensional topological phases

Dimensional evolution between one- ($1D$) and two-dimensional ($2D$) topological phases is investigated systematically. The crossover from a $2D$ topological insulator to its $1D$ limit shows oscillating behavior between a $1D$ ordinary insulator and a $1D$ topological insulator. By constructing a $2D$ topological system from a $1D$ topological insulator, it is shown that there exist possibly weak topological phases in $2D$ time-reversal invariant band insulators, one of which can be realized in anisotropic systems. The topological invariant of the phase is $Z_{2}=0$. However the edge states may appear along specific boundaries. It can be interpreted as arranged $1D$ topological phases, and have symmetry-protecting nature as the corresponding $1D$ topological phase. Robust edge states can exist under specific conditions. These results provide further understanding on $2D$ time-reversal invariant insulators, and can be realized experimentally.

cond-mat.mes-hall↗

arXiv2011

The Optical Counterpart of NGC 1313 X-1

We identify the optical counterpart of the ultraluminous X-ray source (ULX) NGC 1313 X-1 and discuss constraints on its physical nature from multiband optical spectra. There is a single object on Hubble Space Telescope (HST) images within the aspect-corrected Chandra X-ray error circle; a fainter, possibly extended, feature lies near the edge of the error circle. The brighter object showed prominent variation in the F555W band, but was constant in the F814W band. The spectrum was consistent with a single power-law on 2003 Nov 17, but deviated from this on 2004 Jul 17, suggestive of more than one emission component. Based on the location, magnitudes, spectral shape, and variability of the bright object, it is likely the ULX counterpart. The red wing of the spectrum around F814W may be due to emission from the companion star, and the blue wing is likely from disk emission. The stellar population around X-1 has an age older than 30 Myr, without very blue stars or young clusters. This places a constraint on the companion mass of the ULX as no more than 10 solar masses.

astro-ph.HE↗

arXiv2010

Biomimetic isotropic nanostructures for structural coloration

We describe the self-assembly of biomimetic isotropic films which display structural color amenable to potential applications in coatings. Isotropic structures can produce color if there is a pronounced characteristic length-scale comparable to the wavelength of visible light and wavelength-independent scattering is suppressed.

cond-mat.soft↗