Xiang Zhang

Bio: Xiang Zhang is an academic researcher from Baylor College of Medicine. The author has contributed to research in topics: Medicine & Computer science. The author has an hindex of 154, co-authored 1733 publications receiving 117576 citations. Previous affiliations of Xiang Zhang include University of California, Berkeley & University of Texas MD Anderson Cancer Center.

Automated evaluation of building alignments in generalized maps

Abstract: Evaluation is a key step to examine the quality of generalized maps with respect to map requirements. Map generalization facilitates the recognition of pattern generating processes by preserving and highlighting the patterns at smaller scales. This article focuses specifically on the evaluation of building patterns in topographic maps that are generalized from large to mid scales. Currently, there is a lack of knowledge and functionality on automatically evaluating how these patterns are generalized. The issues of the evaluation range from missing formal map requirements on building alignments to missing automated evaluation techniques. This article firstly analyses the requirements constraints related to the generalization of building alignments. Then, it focuses on three more specific constraints, i.e. on existence, orientation of alignments and spatial distribution of composing buildings. Later, a three-step approach is proposed to 1 recognize and 2 match alignments from source and generalized datasets and 3 evaluate building alignments in generalized datasets. Besides, many-to-many and partial matching between initial and target alignments is a side effect of generalization, which reduces the reliability of the evaluation results. This article introduces a confidence indicator to document the reliability and to inform intended users e.g. cartographers and/or systems about the reliability of evaluation decisions. The effectiveness of our approach is demonstrated by evaluating the alignments in both interactively manually generalized maps and automated generalized maps. Finally, we discuss how our approach can be used to control automated generalization and identify further improvements.

Near-field Moiré effect mediated by surface plasmon polariton excitation

Abstract: We have demonstrated a surface plasmon polariton mediated optical Moire effect by inserting a silver slab between two subwavelength gratings. Enhancement of the evanescent fields by the surface plasmon excitations on the silver slab leads to a remarkable contrast improvement in the Moire fringes from two subwavelength gratings. Numerical calculations, which agree very well with the experimental observation of evanescent-wave Moire fringes, elucidate the crucial role of the surface plasmon polaritons. The near-field Moire effect has potential applications to extend the existing Moire techniques to subwavelength characterization of nanostructures.

Super-resolution imaging by random adsorbed molecule probes.

Abstract: Single molecule localization (SML) is a powerful tool to measure the position and trajectory of molecules in numerous systems, with nanometer accuracy. This technique has been recently utilized to overcome the diffraction limit in optical imaging. So far, super-resolution imaging by SML was demonstrated using photoactivable or photoswitchable fluorophores, as well as diffusive fluorophore probes in solution. All these methods, however, rely on special fluorophore or object properties. In this Letter, we propose and demonstrate a new super-resolution technique attainable for a bio/dielectric structure on a metal substrate. A sub-diffraction-limited image is obtained by randomly adsorbed fluorescent probe molecules on a liquid-solid interface, while the metal substrate, quenching the unwanted fluorescent signal, provides a significantly enhanced imaging contrast. As this approach does not use specific stain techniques, it can be readily applied to general dielectric objects, such as nanopatterned photoresist, inorganic nanowires, subcellular structures, etc.

Local Community Detection in Multiple Networks

Abstract: Local community detection aims to find a set of densely-connected nodes containing given query nodes. Most existing local community detection methods are designed for a single network. However, a single network can be noisy and incomplete. Multiple networks are more informative in real-world applications. There are multiple types of nodes and multiple types of node proximities. Complementary information from different networks helps to improve detection accuracy. In this paper, we propose a novel RWM (Random Walk in Multiple networks) model to find relevant local communities in all networks for a given query node set from one network. RWM sends a random walker in each network to obtain the local proximity w.r.t. the query nodes (i.e., node visiting probabilities). Walkers with similar visiting probabilities reinforce each other. They restrict the probability propagation around the query nodes to identify relevant subgraphs in each network and disregard irrelevant parts. We provide rigorous theoretical foundations for RWM and develop two speeding-up strategies with performance guarantees. Comprehensive experiments are conducted on synthetic and real-world datasets to evaluate the effectiveness and efficiency of RWM.

Nanoporous {Y2}-Organic Frameworks for Excellent Catalytic Performance on the Cycloaddition Reaction of Epoxides with CO2 and Deacetalization-Knoevenagel Condensation.

Abstract: Stable metal-organic frameworks containing periodically arranged nanosized pores and active Lewis acid-base active sites are considered as ideal candidates for efficient heterogeneous catalysis. Herein, the exquisite combination of [Y2(CO2)7(H2O)2] cluster (abbreviated as {Y2}) and multifunctional linker of 2,4,6-tri(2,4-dicarboxyphenyl)pyridine (H6TDP) led to a nanoporous framework of {[Y2(TDP)(H2O)2]·5H2O·4DMF}n (NUC-53, NUC = North University of China), which is a rarely reported binuclear three-dimensional (3D) framework with hierarchical tetragonal-microporous (0.78 nm) and octagonal-nanoporous (1.75 nm) channels. The inner walls of these channels are aligned by {Y2} clusters and plentifully coexisted Lewis acid-base sites of YIII ions and Npyridine atoms. Furthermore, NUC-53 has a quite large void volume of ∼65.2%, which is significantly higher than most documented 3D rare-earth-based MOFs. The performed catalytic experiments exhibited that activated NUC-53 showed a high catalytic activity on the cycloaddition reactions of CO2 with styrene oxide under mild conditions with excellent turnover number (TON: 1980) and turnover frequency (TOF: 495 h-1). Moreover, the deacetalization-Knoevenagel condensation reactions of benzaldehyde dimethyl acetal and malononitrile could be efficiently prompted by the heterogeneous catalyst of NUC-53. These findings not only pave the way for the construction of nanoporous MOF based on rare-earth clusters with a variety of catalytic activities but also provide some new insights into the catalytic mechanism.

Deep Residual Learning for Image Recognition

Abstract: Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. On the ImageNet dataset we evaluate residual nets with a depth of up to 152 layers—8× deeper than VGG nets [40] but still having lower complexity. An ensemble of these residual nets achieves 3.57% error on the ImageNet test set. This result won the 1st place on the ILSVRC 2015 classification task. We also present analysis on CIFAR-10 with 100 and 1000 layers. The depth of representations is of central importance for many visual recognition tasks. Solely due to our extremely deep representations, we obtain a 28% relative improvement on the COCO object detection dataset. Deep residual nets are foundations of our submissions to ILSVRC & COCO 2015 competitions1, where we also won the 1st places on the tasks of ImageNet detection, ImageNet localization, COCO detection, and COCO segmentation.

Very Deep Convolutional Networks for Large-Scale Image Recognition

Abstract: In this work we investigate the effect of the convolutional network depth on its accuracy in the large-scale image recognition setting. Our main contribution is a thorough evaluation of networks of increasing depth using an architecture with very small (3x3) convolution filters, which shows that a significant improvement on the prior-art configurations can be achieved by pushing the depth to 16-19 weight layers. These findings were the basis of our ImageNet Challenge 2014 submission, where our team secured the first and the second places in the localisation and classification tracks respectively. We also show that our representations generalise well to other datasets, where they achieve state-of-the-art results. We have made our two best-performing ConvNet models publicly available to facilitate further research on the use of deep visual representations in computer vision.

Deep Residual Learning for Image Recognition

Abstract: Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. On the ImageNet dataset we evaluate residual nets with a depth of up to 152 layers---8x deeper than VGG nets but still having lower complexity. An ensemble of these residual nets achieves 3.57% error on the ImageNet test set. This result won the 1st place on the ILSVRC 2015 classification task. We also present analysis on CIFAR-10 with 100 and 1000 layers. The depth of representations is of central importance for many visual recognition tasks. Solely due to our extremely deep representations, we obtain a 28% relative improvement on the COCO object detection dataset. Deep residual nets are foundations of our submissions to ILSVRC & COCO 2015 competitions, where we also won the 1st places on the tasks of ImageNet detection, ImageNet localization, COCO detection, and COCO segmentation.

Going deeper with convolutions

Abstract: We propose a deep convolutional neural network architecture codenamed Inception that achieves the new state of the art for classification and detection in the ImageNet Large-Scale Visual Recognition Challenge 2014 (ILSVRC14). The main hallmark of this architecture is the improved utilization of the computing resources inside the network. By a carefully crafted design, we increased the depth and width of the network while keeping the computational budget constant. To optimize quality, the architectural decisions were based on the Hebbian principle and the intuition of multi-scale processing. One particular incarnation used in our submission for ILSVRC14 is called GoogLeNet, a 22 layers deep network, the quality of which is assessed in the context of classification and detection.