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.

Metastasis Dormancy in Estrogen Receptor–Positive Breast Cancer

Abstract: About 20% to 40% of patients with breast cancer eventually develop recurrences in distant organs, which are often not detected until years to decades after the primary tumor diagnosis. This phenomenon is especially pronounced in estrogen receptor-positive (ER(+)) breast cancer, suggesting that ER(+) cancer cells may stay dormant for a protracted period of time, despite adjuvant therapies. Multiple mechanisms have been proposed to explain how cancer cells survive and remain in dormancy, and how they become reactivated and exit dormancy. These mechanisms include angiogenic switch, immunosurveillance, and interaction with extracellular matrix and stromal cells. How to eradicate or suppress these dormant cancer cells remains a major clinical issue because of the lack of knowledge about the biologic and clinical nature of these cells. Herein, we review the clinical manifestation of metastasis dormancy in ER(+) tumors, the current biologic insights regarding tumor dormancy obtained from various experimental models, and the clinical challenges to predict, detect, and treat dormant metastases. We also discuss future research directions toward a better understanding of the biologic mechanisms and clinical management of ER(+) dormant metastasis.

High-speed acoustic communication by multiplexing orbital angular momentum.

Abstract: Long-range acoustic communication is crucial to underwater applications such as collection of scientific data from benthic stations, ocean geology, and remote control of off-shore industrial activities. However, the transmission rate of acoustic communication is always limited by the narrow-frequency bandwidth of the acoustic waves because of the large attenuation for high-frequency sound in water. Here, we demonstrate a high-throughput communication approach using the orbital angular momentum (OAM) of acoustic vortex beams with one order enhancement of the data transmission rate at a single frequency. The topological charges of OAM provide intrinsically orthogonal channels, offering a unique ability to multiplex data transmission within a single acoustic beam generated by a transducer array, drastically increasing the information channels and capacity of acoustic communication. A high spectral efficiency of 8.0 ± 0.4 (bit/s)/Hz in acoustic communication has been achieved using topological charges between -4 and +4 without applying other communication modulation techniques. Such OAM is a completely independent degree of freedom which can be readily integrated with other state-of-the-art communication modulation techniques like quadrature amplitude modulation (QAM) and phase-shift keying (PSK). Information multiplexing through OAM opens a dimension for acoustic communication, providing a data transmission rate that is critical for underwater applications.

Novel hollow mesoporous 1D TiO2 nanofibers as photovoltaic and photocatalytic materials.

Abstract: Hollow mesoporous one dimensional (1D) TiO2 nanofibers are successfully prepared by co-axial electrospinning of a titanium tetraisopropoxide (TTIP) solution with two immiscible polymers; polyethylene oxide (PEO) and polyvinylpyrrolidone (PVP) using a core–shell spinneret, followed by annealing at 450 °C. The annealed mesoporous TiO2 nanofibers are found to having a hollow structure with an average diameter of 130 nm. Measurements using the Brunauer–Emmett–Teller (BET) method reveal that hollow mesoporous TiO2 nanofibers possess a high surface area of 118 m2 g−1 with two types of mesopores; 3.2 nm and 5.4 nm that resulted from gaseous removal of PEO and PVP respectively during annealing. With hollow mesoporous TiO2 nanofibers as the photoelectrode in dye sensitized solar cells (DSSC), the solar-to-current conversion efficiency (η) and short circuit current (Jsc) are measured as 5.6% and 10.38 mA cm−2 respectively, which are higher than those of DSSC made using regular TiO2 nanofibers under identical conditions (η = 4.2%, Jsc = 8.99 mA cm−2). The improvement in the conversion efficiency is mainly attributed to the higher surface area and mesoporous TiO2 nanostructure. It facilitates the adsorption of more dye molecules and also promotes the incident photon to electron conversion. Hollow mesoporous TiO2 nanofibers with close packing of grains and crystals intergrown with each other demonstrate faster electron diffusion, and longer electron recombination time than regular TiO2 nanofibers as well as P25 nanoparticles. The surface effect of hollow mesoporous TiO2 nanofibers as a photocatalyst for the degradation of rhodamine dye was also investigated. The kinetic study shows that the hollow mesoporous surface of the TiO2 nanofibers influenced its interactions with the dye, and resulted in an increased catalytic activity over P25 TiO2 nanocatalysts.

High thermoelectric power factor in two-dimensional crystals of Mo S 2

Abstract: The quest for high-efficiency heat-to-electricity conversion has been one of the major driving forces toward renewable energy production for the future. Efficient thermoelectric devices require high voltage generation from a temperature gradient and a large electrical conductivity while maintaining a low thermal conductivity. For a given thermal conductivity and temperature, the thermoelectric power factor is determined by the electronic structure of the material. Low dimensionality (1D and 2D) opens new routes to a high power factor due to the unique density of states (DOS) of confined electrons and holes. The 2D transition metal dichalcogenide (TMDC) semiconductors represent a new class of thermoelectric materials not only due to such confinement effects but especially due to their large effective masses and valley degeneracies. Here, we report a power factor of $\mathrm{Mo}{\mathrm{S}}_{2}$ as large as $8.5\phantom{\rule{0.28em}{0ex}}\mathrm{mW}\phantom{\rule{0.28em}{0ex}}{\mathrm{m}}^{\ensuremath{-}1}\phantom{\rule{0.28em}{0ex}}{\mathrm{K}}^{\ensuremath{-}2}$ at room temperature, which is among the highest measured in traditional, gapped thermoelectric materials. To obtain these high power factors, we perform thermoelectric measurements on few-layer $\mathrm{Mo}{\mathrm{S}}_{2}$ in the metallic regime, which allows us to access the 2D DOS near the conduction band edge and exploit the effect of 2D confinement on electron scattering rates, resulting in a large Seebeck coefficient. The demonstrated high, electronically modulated power factor in 2D TMDCs holds promise for efficient thermoelectric energy conversion.

K∗(892)0 and φ(1020) production in Pb-Pb collisions at sNN =2.76 TeV

Abstract: The yields of the K*(892)(0) and phi(1020) resonances are measured in Pb-Pb collisions at root s(NN) = 2.76 TeV through their hadronic decays using the ALICE detector. The measurements are performed in multiple centrality intervals at mid-rapidity (vertical bar y vertical bar <0.5) in the transverse-momentum ranges 0.3

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.