arXiv subjects

Jie Zhang

Publications and source records attributed to Jie Zhang.

At least 1,009 records · Page 56Linked to original sources

arXiv2019

Regularize, Expand and Compress: Multi-task based Lifelong Learning via NonExpansive AutoML

Lifelong learning, the problem of continual learning where tasks arrive in sequence, has been lately attracting more attention in the computer vision community. The aim of lifelong learning is to develop a system that can learn new tasks while maintaining the performance on the previously learned tasks. However, there are two obstacles for lifelong learning of deep neural networks: catastrophic forgetting and capacity limitation. To solve the above issues, inspired by the recent breakthroughs in automatically learning good neural network architectures, we develop a Multi-task based lifelong learning via nonexpansive AutoML framework termed Regularize, Expand and Compress (REC). REC is composed of three stages: 1) continually learns the sequential tasks without the learned tasks' data via a newly proposed multi-task weight consolidation (MWC) algorithm; 2) expands the network to help the lifelong learning with potentially improved model capability and performance by network-transformation based AutoML; 3) compresses the expanded model after learning every new task to maintain model efficiency and performance. The proposed MWC and REC algorithms achieve superior performance over other lifelong learning algorithms on four different datasets.

cs.CV↗

arXiv2019

FUSE: Fusing STT-MRAM into GPUs to Alleviate Off-Chip Memory Access Overheads

In this work, we propose FUSE, a novel GPU cache system that integrates spin-transfer torque magnetic random-access memory (STT-MRAM) into the on-chip L1D cache. FUSE can minimize the number of outgoing memory accesses over the interconnection network of GPU's multiprocessors, which in turn can considerably improve the level of massive computing parallelism in GPUs. Specifically, FUSE predicts a read-level of GPU memory accesses by extracting GPU runtime information and places write-once-read-multiple (WORM) data blocks into the STT-MRAM, while accommodating write-multiple data blocks over a small portion of SRAM in the L1D cache. To further reduce the off-chip memory accesses, FUSE also allows WORM data blocks to be allocated anywhere in the STT-MRAM by approximating the associativity with the limited number of tag comparators and I/O peripherals. Our evaluation results show that, in comparison to a traditional GPU cache, our proposed heterogeneous cache reduces the number of outgoing memory references by 32% across the interconnection network, thereby improving the overall performance by 217% and reducing energy cost by 53%.

cs.AR↗

arXiv2019

Flux density measurements for 32 pulsars in the 20 cm band

Flux density measurements provide fundamental observational parameters that describe a pulsar. In the current pulsar catalogue, 27% of radio pulsars have no flux density measurement in the 20 cm observing band. Here, we present the first measurements of the flux densities in this band for 32 pulsars observed using the Parkes radio telescope and provide updated pulse profiles for these pulsars. We have used both archival and new observations to make these measurements. Various schemes exist for measuring flux densities. We show how the flux densities measured vary between these methods and how the presence of radio-frequency-interference will bias flux density measurements

astro-ph.GA↗

arXiv2019

Advanced Denoising for X-ray Ptychography

The success of ptychographic imaging experiments strongly depends on achieving high signal-to-noise ratio. This is particularly important in nanoscale imaging experiments when diffraction signals are very weak and the experiments are accompanied by significant parasitic scattering (background), outliers or correlated noise sources. It is also critical when rare events such as cosmic rays, or bad frames caused by electronic glitches or shutter timing malfunction take place. In this paper, we propose a novel iterative algorithm with rigorous analysis that exploits the direct forward model for parasitic noise and sample smoothness to achieve a thorough characterization and removal of structured and random noise. We present a formal description of the proposed algorithm and prove its convergence under mild conditions. Numerical experiments from simulations and real data (both soft and hard X-ray beamlines) demonstrate that the proposed algorithms produce better results when compared to state-of-the-art methods.

eess.IV↗

arXiv2019

Terahertz oscilloscope for recording time information of ultrashort electron beams

We propose and demonstrate a Terahertz (THz) oscilloscope for recording time information of an ultrashort electron beam. By injecting a laser-driven THz pulse with circular polarization into a dielectric tube, the electron beam is swept helically such that the time information is uniformly encoded into the angular distribution that allows one to characterize both the temporal profile and timing jitter of an electron beam. The dynamic range of the measurement in such a configuration is significantly increased compared to deflection with a linearly polarized THz pulse. With this THz oscilloscope, nearly 50-fold longitudinal compression of a relativistic electron beam to about 15 fs (rms) is directly visualized with its arrival time determined with 3 fs accuracy. This technique bridges the gap between streaking of photoelectrons with optical lasers and deflection of relativistic electron beams with radio-frequency deflectors, and should have wide applications in many ultrashort electron beam based facilities.

physics.acc-ph↗

arXiv2019

MICIK: MIning Cross-Layer Inherent Similarity Knowledge for Deep Model Compression

State-of-the-art deep model compression methods exploit the low-rank approximation and sparsity pruning to remove redundant parameters from a learned hidden layer. However, they process each hidden layer individually while neglecting the common components across layers, and thus are not able to fully exploit the potential redundancy space for compression. To solve the above problem and enable further compression of a model, removing the cross-layer redundancy and mining the layer-wise inheritance knowledge is necessary. In this paper, we introduce a holistic model compression framework, namely MIning Cross-layer Inherent similarity Knowledge (MICIK), to fully excavate the potential redundancy space. The proposed MICIK framework simultaneously, (1) learns the common and unique weight components across deep neural network layers to increase compression rate; (2) preserves the inherent similarity knowledge of nearby layers and distant layers to minimize the accuracy loss and (3) can be complementary to other existing compression techniques such as knowledge distillation. Extensive experiments on large-scale convolutional neural networks demonstrate that MICIK is superior over state-of-the-art model compression approaches with 16X parameter reduction on VGG-16 and 6X on GoogLeNet, all without accuracy loss.

cs.CV↗

arXiv2019

Exploring Communities in Large Profiled Graphs

Given a graph $G$ and a vertex $q\in G$, the community search (CS) problem aims to efficiently find a subgraph of $G$ whose vertices are closely related to $q$. Communities are prevalent in social and biological networks, and can be used in product advertisement and social event recommendation. In this paper, we study profiled community search (PCS), where CS is performed on a profiled graph. This is a graph in which each vertex has labels arranged in a hierarchical manner. Extensive experiments show that PCS can identify communities with themes that are common to their vertices, and is more effective than existing CS approaches. As a naive solution for PCS is highly expensive, we have also developed a tree index, which facilitate efficient and online solutions for PCS.

cs.DB↗

arXiv2019

The Origin of Major Solar Activity - Collisional Shearing Between Nonconjugated Polarities of Multiple Bipoles Emerging Within Active Regions

Active Regions (ARs) that exhibit compact Polarity Inversion Lines (PILs) are known to be very flare-productive. However, the physical mechanisms behind this statistical inference have not been demonstrated conclusively. We show that such PILs can occur due to the collision between two emerging flux tubes nested within the same AR. In such multipolar ARs, the flux tubes may emerge simultaneously or sequentially, each initially producing a bipolar magnetic region (BMR) at the surface. During each flux tube's emergence phase, the magnetic polarities can migrate such that opposite polarities belonging to different BMRs collide, resulting in shearing and cancellation of magnetic flux. We name this process 'collisional shearing' to emphasize that the shearing and flux cancellation develops due to the collision. Collisional shearing is a process different from the known concept of flux cancellation occurring between polarities of a single bipole, a process that has been commonly used in many numerical models. High spatial and temporal resolution observations from the Solar Dynamics Observatory for two emerging ARs, AR11158 and AR12017, show the continuous cancellation of up to 40% of the unsigned magnetic flux of the smallest BMR, which occurs at the collisional PIL for as long as the collision persists. The flux cancellation is accompanied by a succession of solar flares and CMEs, products of magnetic reconnection along the collisional PIL. Our results suggest that the quantification of magnetic cancellation driven by collisional shearing needs to be taken into consideration in order to improve the prediction of solar energetic events and space weather.

astro-ph.SR↗

arXiv2018

A hybrid silicon-sapphire cryogenic Fabry-Perot cavity using hydroxide catalysis bonding

The third-generation gravitational wave detectors are under development by operating the detector in cryogenic temperature to reduce the thermal noise. Silicon and sapphire are promising candidate materials for the test masses and suspension elements due to their remarkable mechanical and thermal properties at cryogenic temperature. Here we present the performances of the cryogenic thermal cycling and strength testing on hydroxide catalysis bonding between sapphire and silicon. Our results suggest that although these two materials have very different coefficients of thermal expansion, but if the flatness and the thermally grown $\mathrm{SiO_2}$ oxidation layer on the silicon surface are controlled well, the bonded samples can still survive thermal cycling from room temperature to 5.5 K. A breaking strength of 3.6$\pm 0.6$ MPa is measured for the bonds between sapphire and silicon with a 190 nm silicon oxidation thickness after cooling cycle. We construct a hybrid sapphire-silicon Fabry-Perot cavity with the developing bonding technique in our lab. The measurement results reveal that the cavity can survive repeated thermal cycling while maintaining a good finesse.

physics.ins-det↗

arXiv2018

Disentangling boson peaks and Van Hove singularities in a model glass

Using the example of a two-dimensional macroscopic model glass in which the interparticle forces can be precisely measured, we obtain strong hints for resolving a controversy concerning the origin of the anomalous enhancement of the vibrational spectrum in glasses (boson peak). Whereas many authors attribute this anomaly to the structural disorder, some other authors claim that the short-range order, leading to washed-out Van Hove singularities, would cause the boson-peak anomaly. As in our model system, the disorder-induced and shortrange--order-induced features can be completely separated, we are able to discuss the controversy about the boson peak in real glasses in a new light. Our findings suggest that the interpretation of the boson peak in terms of short-range order only, might result from a coincidence of the two phenomena in the materials studied. In general, as we show, the two phenomena both exist, but are two completely separate entities.

cond-mat.soft↗

arXiv2018

Localized polarons and conductive charge carriers: understanding CaCu$_{3}$Ti$_{4}$O$_{12}$ over a broad temperature range

CaCu$_{3}$Ti$_{4}$O$_{12}$ (CCTO) has a large dielectric permittivity that is independent of the probing frequency near the room temperature, which complicated due to the existence of several dynamic processes. Here, we consider the combined effects of localized charge carriers (polarons) and thermally activated charge carriers using a recently proposed statistical model to fit and understand the permittivity of CCTO measured at different frequencies over the whole temperature range accessible by our experiments. We found that the small permittivity at the lowest temperature is related to polaron frozen, while at higher temperatures the rapid increase is associated with the thermal excitation of polarons inducing the Maxwell-Wagner effect, and the final increase of the permittivity is attributed to the thermally activated conductivity. Such analysis enables us to separate the contributions from localized polarons and conductive charge carriers and quantify their activation energies.

cond-mat.mtrl-sci↗

arXiv2018

Amber: Enabling Precise Full-System Simulation with Detailed Modeling of All SSD Resources

SSDs become a major storage component in modern memory hierarchies, and SSD research demands exploring future simulation-based studies by integrating SSD subsystems into a full-system environment. However, several challenges exist to model SSDs under a full-system simulations; SSDs are composed upon their own complete system and architecture, which employ all necessary hardware, such as CPUs, DRAM and interconnect network. Employing the hardware components, SSDs also require to have multiple device controllers, internal caches and software modules that respect a wide spectrum of storage interfaces and protocols. These SSD hardware and software are all necessary to incarnate storage subsystems under full-system environment, which can operate in parallel with the host system. In this work, we introduce a new SSD simulation framework, SimpleSSD 2.0, namely Amber, that models embedded CPU cores, DRAMs, and various flash technologies (within an SSD), and operate under the full system simulation environment by enabling a data transfer emulation. Amber also includes full firmware stack, including DRAM cache logic, flash firmware, such as FTL and HIL, and obey diverse standard protocols by revising the host DMA engines and system buses of a popular full system simulator's all functional and timing CPU models (gem5). The proposed simulator can capture the details of dynamic performance and power of embedded cores, DRAMs, firmware and flash under the executions of various OS systems and hardware platforms. Using Amber, we characterize several system-level challenges by simulating different types of fullsystems, such as mobile devices and general-purpose computers, and offer comprehensive analyses by comparing passive storage and active storage architectures.

cs.AR↗

arXiv2018

Reinforcement Learning based Dynamic Model Selection for Short-Term Load Forecasting

With the growing prevalence of smart grid technology, short-term load forecasting (STLF) becomes particularly important in power system operations. There is a large collection of methods developed for STLF, but selecting a suitable method under varying conditions is still challenging. This paper develops a novel reinforcement learning based dynamic model selection (DMS) method for STLF. A forecasting model pool is first built, including ten state-of-the-art machine learning based forecasting models. Then a Q-learning agent learns the optimal policy of selecting the best forecasting model for the next time step, based on the model performance. The optimal DMS policy is applied to select the best model at each time step with a moving window. Numerical simulations on two-year load and weather data show that the Q-learning algorithm converges fast, resulting in effective and efficient DMS. The developed STLF model with Q-learning based DMS improves the forecasting accuracy by approximately 50%, compared to the state-of-the-art machine learning based STLF models.

cs.LG↗

arXiv2018

Roadmap for Reliable Ensemble Forecasting of the Sun-Earth System

The authors of this report met on 28-30 March 2018 at the New Jersey Institute of Technology, Newark, New Jersey, for a 3-day workshop that brought together a group of data providers, expert modelers, and computer and data scientists, in the solar discipline. Their objective was to identify challenges in the path towards building an effective framework to achieve transformative advances in the understanding and forecasting of the Sun-Earth system from the upper convection zone of the Sun to the Earth's magnetosphere. The workshop aimed to develop a research roadmap that targets the scientific challenge of coupling observations and modeling with emerging data-science research to extract knowledge from the large volumes of data (observed and simulated) while stimulating computer science with new research applications. The desire among the attendees was to promote future trans-disciplinary collaborations and identify areas of convergence across disciplines. The workshop combined a set of plenary sessions featuring invited introductory talks and workshop progress reports, interleaved with a set of breakout sessions focused on specific topics of interest. Each breakout group generated short documents, listing the challenges identified during their discussions in addition to possible ways of attacking them collectively. These documents were combined into this report-wherein a list of prioritized activities have been collated, shared and endorsed.

astro-ph.SR↗

arXiv2018

Separating the effects of experimental noise from inherent system variability in voltammetry: the $[$Fe(CN)$_6]^{3-/ 4-}$ process

Recently, we have introduced the use of techniques drawn from Bayesian statistics to recover kinetic and thermodynamic parameters from voltammetric data, and were able to show that the technique of large amplitude ac voltammetry yielded significantly more accurate parameter values than the equivalent dc approach. In this paper we build on this work to show that this approach allows us, for the first time, to separate the effects of random experimental noise and inherent system variability in voltammetric experiments. We analyse ten repeated experimental data sets for the $[$Fe(CN)$_6]^{3-/ 4-}$ process, again using large-amplitude ac cyclic voltammetry. In each of the ten cases we are able to obtain an extremely good fit to the experimental data and obtain very narrow distributions of the recovered parameters governing both the faradaic (the reversible formal faradaic potential, $E_0$, the standard heterogeneous charge transfer rate constant $k_0$, and the charge transfer coefficient $α$) and non-faradaic terms (uncompensated resistance, $R_u$, and double layer capacitance, $C_{dl}$). We then employ hierarchical Bayesian methods to recover the underlying "hyperdistribution" of the faradaic and non-faradaic parameters, showing that in general the variation between the experimental data sets is significantly greater than suggested by individual experiments, except for $α$ where the inter-experiment variation was relatively minor. Correlations between pairs of parameters are provided, and for example, reveal a weak link between $k_0$ and $C_{dl}$ (surface activity of a glassy carbon electrode surface). Finally, we discuss the implications of our findings for voltammetric experiments more generally.

eess.SP↗

arXiv2018

Equitable Partition Theorem of Tensors and Spectrum of Generalized Power Hypergraphs

In this paper, we present an equitable partition theorem of tensors, which gives the relations between $H$-eigenvalues of a tensor and its quotient equitable tensor and extends the equitable partitions of graphs to hypergraphs. Furthermore, with the aid of it, some properties and $H$-eigenvalues of the generalized power hypergraphs are obtained, which extends some known results, including some results of Yuan, Qi and Shao.

math.CO↗

arXiv2018

The credibility of Rydberg atom based digital communication over a continuously tunable radio-frequency carrier

High-sensitive measurement of radio-frequency (RF) electric field is available via the electromagnetically induced transparency (EIT) effect of Rydberg atom at room-temperature, which has been developed to be a promising atomic RF receiver. In this Letter, we investigate the credibility of the digital communication via this quantum-based antenna over the entire continuously tunable RF-carrier. Our experiment shows that digital communication at a rate of 500 kbps performs reliably within a tunable bandwidth of 200 MHz at carrier 10.22 GHz and a bit error rate (BER) appears out of this range, for example, the BER runs up to 15 % at RF-detuning $\pm150$ MHz. In the measurement, the time-variant RF field is retrieved by detecting the density of the probe laser at the center frequency of RF-induced symmetric or asymmetric Autler-Townes splitting in EIT. Prior to the digital test, we have studied the RF-receiving quality versus the physical ambiance and found that a choice of linear gain response to the RF-amplitude can suppress the signal distortion and the modulating signal is able to be decoded as fast as up to 500 kHz in the tunable bandwidth. Our checkout consolidates the physical foundation for a reliable communication and spectrum sensing over the broadband RF-carrier.

physics.atom-ph↗

arXiv2018

Semi-supervised Learning on Graphs with Generative Adversarial Nets

We investigate how generative adversarial nets (GANs) can help semi-supervised learning on graphs. We first provide insights on working principles of adversarial learning over graphs and then present GraphSGAN, a novel approach to semi-supervised learning on graphs. In GraphSGAN, generator and classifier networks play a novel competitive game. At equilibrium, generator generates fake samples in low-density areas between subgraphs. In order to discriminate fake samples from the real, classifier implicitly takes the density property of subgraph into consideration. An efficient adversarial learning algorithm has been developed to improve traditional normalized graph Laplacian regularization with a theoretical guarantee. Experimental results on several different genres of datasets show that the proposed GraphSGAN significantly outperforms several state-of-the-art methods. GraphSGAN can be also trained using mini-batch, thus enjoys the scalability advantage.

cs.SI↗