arXiv subjects

Cheng Chen

Publications and source records attributed to Cheng Chen.

At least 379 records · Page 21Linked to original sources

arXiv2018

Detection of bosonic mode as a signature of magnetic excitation in one unit cell FeSe on SrTiO3

We report an in situ scanning tunneling spectroscopy study of one-unit-cell (1-UC) FeSe film on SrTiO3(001) (STO) substrate. In quasiparticle density of states, bosonic excitation mode characterized by the "dip-hump" structure is detected outside the larger superconducting gap with energy comparable with phonon and spin resonance modes in heavily electron-doped iron selenides. Statistically, the excitation mode, which is intimately correlated with superconductivity, shows an anticorrelation with pairing strength and yields an energy scale upper-bounded by twice the superconducting gap coinciding with the characteristics of magnetic resonance in cuprates and iron-based superconductors. The local response of tunneling spectra to magnetically different Se defects all exhibits the induced in-gap quasiparticle bound states, indicating an unconventional sign-reversing pairing. These results support the magnetic nature of the excitation mode and possibly reveal a signature of electron-magnetic-excitation coupling in high-temperature superconductivity of 1-UC FeSe/STO.

cond-mat.supr-con↗

arXiv2018

Unsupervised Cross-Modality Domain Adaptation of ConvNets for Biomedical Image Segmentations with Adversarial Loss

Convolutional networks (ConvNets) have achieved great successes in various challenging vision tasks. However, the performance of ConvNets would degrade when encountering the domain shift. The domain adaptation is more significant while challenging in the field of biomedical image analysis, where cross-modality data have largely different distributions. Given that annotating the medical data is especially expensive, the supervised transfer learning approaches are not quite optimal. In this paper, we propose an unsupervised domain adaptation framework with adversarial learning for cross-modality biomedical image segmentations. Specifically, our model is based on a dilated fully convolutional network for pixel-wise prediction. Moreover, we build a plug-and-play domain adaptation module (DAM) to map the target input to features which are aligned with source domain feature space. A domain critic module (DCM) is set up for discriminating the feature space of both domains. We optimize the DAM and DCM via an adversarial loss without using any target domain label. Our proposed method is validated by adapting a ConvNet trained with MRI images to unpaired CT data for cardiac structures segmentations, and achieved very promising results.

cs.CV↗

arXiv2018

Semantic-Aware Generative Adversarial Nets for Unsupervised Domain Adaptation in Chest X-ray Segmentation

In spite of the compelling achievements that deep neural networks (DNNs) have made in medical image computing, these deep models often suffer from degraded performance when being applied to new test datasets with domain shift. In this paper, we present a novel unsupervised domain adaptation approach for segmentation tasks by designing semantic-aware generative adversarial networks (GANs). Specifically, we transform the test image into the appearance of source domain, with the semantic structural information being well preserved, which is achieved by imposing a nested adversarial learning in semantic label space. In this way, the segmentation DNN learned from the source domain is able to be directly generalized to the transformed test image, eliminating the need of training a new model for every new target dataset. Our domain adaptation procedure is unsupervised, without using any target domain labels. The adversarial learning of our network is guided by a GAN loss for mapping data distributions, a cycle-consistency loss for retaining pixel-level content, and a semantic-aware loss for enhancing structural information. We validated our method on two different chest X-ray public datasets for left/right lung segmentation. Experimental results show that the segmentation performance of our unsupervised approach is highly competitive with the upper bound of supervised transfer learning.

cs.CV↗

arXiv2018

Design of a Non-vacuum-cooling Compact Scientific CCD Camera

CCD was born in Bell Laboratories in 1969 and has been widely used in various fields. Its ultra-low noise and high quantum efficiency make it work well in particle physics, high energy physics, nuclear physics and astrophysics. Nowadays, more and more CCD cameras have been developed for medical diagnosis, scientific experiments, aerospace, military exploration and other fields. For the wide range of CCD cameras, a Non-vacuum-cooling compact (NVCC) scientific CCD camera has been developed, including FPGA-based low noise clock and bias driver circuit, data acquisition circuit, STM32-based temperature control design. At the same time, the readout noise of the imaging system is studied emphatically. The scheme to generate the CCD clock and the bias driving circuit through ultralow noise LDOs is proposed. The camera was tested in a variety of environments, and the test results show that the system can run at a maximum rate of 5M pixels/s and readout noise is as low as 9.29e^- when the CCD readout speed is 500K pixels/s. Finally, a series of stability tests were carried out on the camera system.

physics.ins-det↗

arXiv2018

RPC Considered Harmful: Fast Distributed Deep Learning on RDMA

Deep learning emerges as an important new resource-intensive workload and has been successfully applied in computer vision, speech, natural language processing, and so on. Distributed deep learning is becoming a necessity to cope with growing data and model sizes. Its computation is typically characterized by a simple tensor data abstraction to model multi-dimensional matrices, a data-flow graph to model computation, and iterative executions with relatively frequent synchronizations, thereby making it substantially different from Map/Reduce style distributed big data computation. RPC, commonly used as the communication primitive, has been adopted by popular deep learning frameworks such as TensorFlow, which uses gRPC. We show that RPC is sub-optimal for distributed deep learning computation, especially on an RDMA-capable network. The tensor abstraction and data-flow graph, coupled with an RDMA network, offers the opportunity to reduce the unnecessary overhead (e.g., memory copy) without sacrificing programmability and generality. In particular, from a data access point of view, a remote machine is abstracted just as a "device" on an RDMA channel, with a simple memory interface for allocating, reading, and writing memory regions. Our graph analyzer looks at both the data flow graph and the tensors to optimize memory allocation and remote data access using this interface. The result is up to 25 times speedup in representative deep learning benchmarks against the standard gRPC in TensorFlow and up to 169% improvement even against an RPC implementation optimized for RDMA, leading to faster convergence in the training process.

cs.DC↗

arXiv2018

Spatial mapping and analysis of aerosols during a forest fire using computational mobile microscopy

Forest fires are a major source of particulate matter (PM) air pollution on a global scale. The composition and impact of PM are typically studied using only laboratory instruments and extrapolated to real fire events owing to a lack of analytical techniques suitable for field-settings. To address this and similar field test challenges, we developed a mobile-microscopy and machine-learning-based air quality monitoring platform called c-Air, which can perform air sampling and microscopic analysis of aerosols in an integrated portable device. We tested its performance for PM sizing and morphological analysis during a recent forest fire event in La Tuna Canyon Park by spatially mapping the PM. The result shows that with decreasing distance to the fire site, the PM concentration increases dramatically, especially for particles smaller than 2 microns. Image analysis from the c-Air portable device also shows that the increased PM is comparatively strongly absorbing and asymmetric, with an aspect ratio of 0.5-0.7. These PM features indicate that a major portion of the PM may be open-flame-combustion-generated element carbon soot-type particles. This initial small-scale experiment shows that c-Air has some potential for forest fire monitoring.

physics.ins-det↗

arXiv2018

GARDENIA: A Domain-specific Benchmark Suite for Next-generation Accelerators

This paper presents the Graph Analytics Repository for Designing Next-generation Accelerators (GARDENIA), a benchmark suite for studying irregular algorithms on massively parallel accelerators. Existing generic benchmarks for accelerators have mainly focused on high performance computing (HPC) applications with limited control and data irregularity, while available graph analytics benchmarks do not apply state-of-the-art algorithms and/or optimization techniques. GARDENIA includes emerging irregular applications in big-data and machine learning domains which mimic massively multithreaded commercial programs running on modern large-scale datacenters. Our characterization shows that GARDENIA exhibits irregular microarchitectural behavior which is quite different from structured workloads and straightforward-implemented graph benchmarks.

cs.PF↗

arXiv2017

Fast simulation of the CEPC detector with Delphes

Fast simulation tools are highly appreciated in particle physics phenomenology studies, especially in the exploration of the physics potential of future experimental facilities. The Circular Electron Positron Collider is a proposed Higgs and Z factory that can precisely measure the Higgs boson properties and the electroweak precision observables. A fast-simulation toolkit dedicated to the CEPC detector has been developed using Delphes. The comparison shows that this fast simulation tool is highly consistent with the full simulation, on a set of benchmark distributions. Therefore, we recommend this fast simulation toolkit for CEPC phenomenological investigations.

hep-ex↗

arXiv2017

Ultrafast, highly-sensitive infrared photodetectors based on two-dimensional oxyselenide crystals

Infrared detection and sensing is deeply embedded in modern technology and human society and its development has always been benefitting from the discovery of new photoelectric response materials. The rise of two-dimensional (2D) materials, thanks to their distinct electronic structure, extreme dimensional confinement and strong light-matter interactions, provides new material platform for next-generation infrared photodetection. Ideal infrared detectors should have fast respond, high sensitivity and air-stability, which is rare to meet at the same time for all existing 2D materials, either graphene, transition metal dichalcogenide or black phosphorous. Herein we demonstrate a new infrared photodetector based on 2D Bi2O2Se crystals, whose main characteristics are superb in the whole 2D family: high sensitivity of ~65 A/W at 1200 nm and ultrafast intrinsic photoresponse of ~1 ps at room temperature. Such great performance is attributed to the suitable electronic bandgap and high carrier mobility of 2D oxyselenide. With additional merits of mass production, excellent stability and flexibility, 2D oxyselenide detectors should open new avenues in highly-sensitive, high-speed, low-cost, flexible infrared photodetection and imaging.

cond-mat.mtrl-sci↗

arXiv2017

$H \rightarrow e^+ e^- $ at CEPC: ISR effect with MadGraph

The Circular Electron Positron Collider (CEPC) is a future Higgs factory proposed by the Chinese high energy physics community. It will operate at a center-of-mass energy of 240-250 GeV. The CEPC will accumulate an integrated luminosity of 5 ab$^{\rm{-1}}$ in ten years' operation. With GEANT4-based full simulation samples for CEPC, Higgs boson decaying into electron pair is studied at the CEPC. The upper limit of ${\cal B}(H \rightarrow e^+ e^-)$ could reach 0.024\% at 95\% confidence level. The signal process is generated by MadGraph, with Initial State Radiation (ISR) implemented, as a first step to adjust MadGraph for a electron positron Collider.

hep-ph↗

arXiv2017

Observation of "broad" d-wave Feshbach resonances with a triplet structure

High partial-wave ($l\ge2$) Feshbach resonance (FR) in an ultracold mixture of $^{85}$Rb-$^{87}$Rb atoms is investigated experimentally aided by a partial-wave insensitive analytic multichannel quantum-defect theory (MQDT). Two "broad" resonances from coupling between d-waves in both the open and closed channels are observed and characterized. One of them shows a fully resolved triplet structure with splitting ratio well explained by the perturbation to the closed channel due to interatomic spin-spin interaction. These tunable "broad" d-wave resonances, especially the one in the lowest-energy open channel, could find important applications in simulating d-wave coupling dominated many-body systems. In addition, we find that there is generally a time and temperature requirement, associated with tunneling through the angular momentum barrier, to establish and observe resonant coupling in nonzero partial waves.

cond-mat.quant-gas↗

arXiv2017

Hybrid Deep-Semantic Matrix Factorization for Tag-Aware Personalized Recommendation

Matrix factorization has now become a dominant solution for personalized recommendation on the Social Web. To alleviate the cold start problem, previous approaches have incorporated various additional sources of information into traditional matrix factorization models. These upgraded models, however, achieve only "marginal" enhancements on the performance of personalized recommendation. Therefore, inspired by the recent development of deep-semantic modeling, we propose a hybrid deep-semantic matrix factorization (HDMF) model to further improve the performance of tag-aware personalized recommendation by integrating the techniques of deep-semantic modeling, hybrid learning, and matrix factorization. Experimental results show that HDMF significantly outperforms the state-of-the-art baselines in tag-aware personalized recommendation, in terms of all evaluation metrics, e.g., its mean reciprocal rank (resp., mean average precision) is 1.52 (resp., 1.66) times as high as that of the best baseline.

cs.IR↗

arXiv2017

Competitive Resource Allocation in HetNets: the Impact of Small-cell Spectrum Constraints and Investment Costs

Heterogeneous wireless networks with small-cell deployments in licensed and unlicensed spectrum bands are a promising approach for expanding wireless connectivity and service. As a result, wireless service providers (SPs) are adding small-cells to augment their existing macro-cell deployments. This added flexibility complicates network management, in particular, service pricing and spectrum allocations across macro- and small-cells. Further, these decisions depend on the degree of competition among SPs. Restrictions on shared spectrum access imposed by regulators, such as low power constraints that lead to small-cell deployments, along with the investment cost needed to add small cells to an existing network, also impact strategic decisions and market efficiency. If the revenue generated by small-cells does not cover the investment cost, then there will be no deployment even if it increases social welfare. We study the implications of such spectrum constraints and investment costs on resource allocation and pricing decisions by competitive SPs, along with the associated social welfare. Our results show that while the optimal resource allocation taking constraints and investment into account can be uniquely determined, adding those features with strategic SPs can have a substantial effect on the equilibrium market structure.

cs.NI↗

arXiv2017

The Impact of Small-Cell Bandwidth Requirements on Strategic Operators

Small-cell deployment in licensed and unlicensed spectrum is considered to be one of the key approaches to cope with the ongoing wireless data demand explosion. Compared to traditional cellular base stations with large transmission power, small-cells typically have relatively low transmission power, which makes them attractive for some spectrum bands that have strict power regulations, for example, the 3.5GHz band [1]. In this paper we consider a heterogeneous wireless network consisting of one or more service providers (SPs). Each SP operates in both macro-cells and small-cells, and provides service to two types of users: mobile and fixed. Mobile users can only associate with macro-cells whereas fixed users can connect to either macro- or small-cells. The SP charges a price per unit rate for each type of service. Each SP is given a fixed amount of bandwidth and splits it between macro- and small-cells. Motivated by bandwidth regulations, such as those for the 3.5Gz band, we assume a minimum amount of bandwidth has to be set aside for small-cells. We study the optimal pricing and bandwidth allocation strategies in both monopoly and competitive scenarios. In the monopoly scenario the strategy is unique. In the competitive scenario there exists a unique Nash equilibrium, which depends on the regulatory constraints. We also analyze the social welfare achieved, and compare it to that without the small-cell bandwidth constraints. Finally, we discuss implications of our results on the effectiveness of the minimum bandwidth constraint on influencing small-cell deployments.

cs.IT↗

arXiv2016

System Design of Internet-of-Things for Residential Smart Grid

Internet-of-Things (IoTs) envisions to integrate, coordinate, communicate, and collaborate real-world objects in order to perform daily tasks in a more intelligent and efficient manner. To comprehend this vision, this paper studies the design of a large scale IoT system for smart grid application, which constitutes a large number of home users and has the requirement of fast response time. In particular, we focus on the messaging protocol of a universal IoT home gateway, where our cloud enabled system consists of a backend server, unified home gateway (UHG) at the end users, and user interface for mobile devices. We discuss the features of such IoT system to support a large scale deployment with a UHG and real-time residential smart grid applications. Based on the requirements, we design an IoT system using the XMPP protocol, and implemented in a testbed for energy management applications. To show the effectiveness of the designed testbed, we present some results using the proposed IoT architecture.

cs.NI↗

arXiv2016

A Novel Method to Study Bottom-up Visual Saliency and its Neural Mechanism

In this study, we propose a novel method to measure bottom-up saliency maps of natural images. In order to eliminate the influence of top-down signals, backward masking is used to make stimuli (natural images) subjectively invisible to subjects, however, the bottom-up saliency can still orient the subjects attention. To measure this orientation/attention effect, we adopt the cueing effect paradigm by deploying discrimination tasks at each location of an image, and measure the discrimination performance variation across the image as the attentional effect of the bottom-up saliency. Such attentional effects are combined to construct a final bottomup saliency map. Based on the proposed method, we introduce a new bottom-up saliency map dataset of natural images to benchmark computational models. We compare several state-of-the-art saliency models on the dataset. Moreover, the proposed paradigm is applied to investigate the neural basis of the bottom-up visual saliency map by analyzing psychophysical and fMRI experimental results. Our findings suggest that the bottom-up saliency maps of natural images are constructed in V1. It provides a strong scientific evidence to resolve the long standing dispute in neuroscience about where the bottom-up saliency map is constructed in human brain.

cs.CV↗

arXiv2016

Evaluating the Performance Impact of Multiple Streams on the MIC-based Heterogeneous Platform

Using \textit{multiple streams} can improve the overall system performance by mitigating the data transfer overhead on heterogeneous systems. Prior work focuses a lot on GPUs but little is known about the performance impact on (Intel Xeon) Phi. In this work, we apply multiple streams into six real-world applications on Phi. We then systematically evaluate the performance benefits of using multiple streams. The evaluation work is performed at two levels: the microbenchmarking level and the real-world application level. Our experimental results at the microbenchmark level show that data transfers and kernel execution can be overlapped on Phi, while data transfers in both directions are performed in a serial manner. At the real-world application level, we show that both overlappable and non-overlappable applications can benefit from using multiple streams (with an performance improvement of up to 24\%). We also quantify how task granularity and resource granularity impact the overall performance. Finally, we present a set of heuristics to reduce the search space when determining a proper task granularity and resource granularity. To conclude, our evaluation work provides lots of insights for runtime and architecture designers when using multiple streams on Phi.

cs.DC↗

arXiv2016

The Impact of Unlicensed Access on Small-Cell Resource Allocation

Small cells deployed in licensed spectrum and unlicensed access via WiFi provide different ways of expanding wireless services to low mobility users. That reduces the demand for conventional macro-cellular networks, which are better suited for wide-area mobile coverage. The mix of these technologies seen in practice depends in part on the decisions made by wireless service providers that seek to maximize revenue, and allocations of licensed and unlicensed spectrum by regulators. To understand these interactions we present a model in which a service provider allocates available licensed spectrum across two separate bands, one for macro- and one for small-cells, in order to serve two types of users: mobile and fixed. We assume a service model in which the providers can charge a (different) price per unit rate for each type of service (macro- or small-cell); unlicensed access is free. With this setup we study how the addition of unlicensed spectrum affects prices and the optimal allocation of bandwidth across macro-/small-cells. We also characterize the optimal fraction of unlicensed spectrum when new bandwidth becomes available.

cs.IT↗