arXiv subjects

Yu Wang

Publications and source records attributed to Yu Wang.

At least 1,279 records · Page 71Linked to original sources

arXiv2018

A New Concept of Deep Reinforcement Learning based Augmented General Sequence Tagging System

In this paper, a new deep reinforcement learning based augmented general sequence tagging system is proposed. The new system contains two parts: a deep neural network (DNN) based sequence tagging model and a deep reinforcement learning (DRL) based augmented tagger. The augmented tagger helps improve system performance by modeling the data with minority tags. The new system is evaluated on SLU and NLU sequence tagging tasks using ATIS and CoNLL-2003 benchmark datasets, to demonstrate the new system's outstanding performance on general tagging tasks. Evaluated by F1 scores, it shows that the new system outperforms the current state-of-the-art model on ATIS dataset by 1.9% and that on CoNLL-2003 dataset by 1.4%.

cs.LG↗

arXiv2018

A Bi-model based RNN Semantic Frame Parsing Model for Intent Detection and Slot Filling

Intent detection and slot filling are two main tasks for building a spoken language understanding(SLU) system. Multiple deep learning based models have demonstrated good results on these tasks . The most effective algorithms are based on the structures of sequence to sequence models (or "encoder-decoder" models), and generate the intents and semantic tags either using separate models or a joint model. Most of the previous studies, however, either treat the intent detection and slot filling as two separate parallel tasks, or use a sequence to sequence model to generate both semantic tags and intent. Most of these approaches use one (joint) NN based model (including encoder-decoder structure) to model two tasks, hence may not fully take advantage of the cross-impact between them. In this paper, new Bi-model based RNN semantic frame parsing network structures are designed to perform the intent detection and slot filling tasks jointly, by considering their cross-impact to each other using two correlated bidirectional LSTMs (BLSTM). Our Bi-model structure with a decoder achieves state-of-the-art result on the benchmark ATIS data, with about 0.5$\%$ intent accuracy improvement and 0.9 $\%$ slot filling improvement.

cs.CL↗

arXiv2018

A Large Catalogue of Multi-wavelength GRB Afterglows I: Color Evolution And Its Physical Implication

The spectrum of gamma-ray burst (GRB) afterglows can be studied with color indices. Here we present a large comprehensive catalogue of 70 GRBs with multi-wavelength optical transient data on which we perform a systematic study to find the temporal evolution of color indices. We categorize them into two samples based on how well the color indices are evaluated. The Golden sample includes 25 bursts mostly observed by GROND, and the Silver sample includes 45 bursts observed by other telescopes. For the Golden sample, we find that 95\% of the color indices do not vary over time. However, the color indices do vary during short periods in most bursts. The observed variations are consistent with effects of (i) the cooling frequency crossing the studied energy bands in a wind medium (43\%) and in a constant density medium (30\%), (ii) early dust extinction (12\%), (iii) transition from reverse shock to forward shock emission (5\%), or (iv) an emergent supernova emission (10\%). We also study the evolutionary properties of the mean color indices for different emission episodes. We find that 86\% of the color indices in the 70 bursts show constancy between consecutive ones. The color index variations occur mainly during the late GRB-SN bump, the flare and early reversed-shock emission components. We further perform a statistical analysis of various observational properties and model parameters (spectral index $β_{o}^{\rm CI}$, electron spectral indices $p^{\rm CI}$, etc.) using color indices. Overall, we conclude that $\sim$ 90\% of colors are constant in time and can be accounted for by the simplest external forward shock model, while the varying color indices call for more detailed modeling.

astro-ph.HE↗

arXiv2018

Hu-Fu: Hardware and Software Collaborative Attack Framework against Neural Networks

Recently, Deep Learning (DL), especially Convolutional Neural Network (CNN), develops rapidly and is applied to many tasks, such as image classification, face recognition, image segmentation, and human detection. Due to its superior performance, DL-based models have a wide range of application in many areas, some of which are extremely safety-critical, e.g. intelligent surveillance and autonomous driving. Due to the latency and privacy problem of cloud computing, embedded accelerators are popular in these safety-critical areas. However, the robustness of the embedded DL system might be harmed by inserting hardware/software Trojans into the accelerator and the neural network model, since the accelerator and deploy tool (or neural network model) are usually provided by third-party companies. Fortunately, inserting hardware Trojans can only achieve inflexible attack, which means that hardware Trojans can easily break down the whole system or exchange two outputs, but can't make CNN recognize unknown pictures as targets. Though inserting software Trojans has more freedom of attack, it often requires tampering input images, which is not easy for attackers. So, in this paper, we propose a hardware-software collaborative attack framework to inject hidden neural network Trojans, which works as a back-door without requiring manipulating input images and is flexible for different scenarios. We test our attack framework for image classification and face recognition tasks, and get attack success rate of 92.6% and 100% on CIFAR10 and YouTube Faces, respectively, while keeping almost the same accuracy as the unattacked model in the normal mode. In addition, we show a specific attack scenario in which a face recognition system is attacked and gives a specific wrong answer.

cs.CR↗

arXiv2018

Crowdsensing Game with Demand Uncertainties: A Deep Reinforcement Learning Approach

Currently, explosive increase of smartphones with powerful built-in sensors such as GPS, accelerometers, gyroscopes and cameras has made the design of crowdsensing applications possible, which create a new interface between human beings and life environment. Until now, various mobile crowdsensing applications have been designed, where the crowdsourcers can employ mobile users (MUs) to complete the required sensing tasks. In this paper, emerging learning-based techniques are leveraged to address crowdsensing game with demand uncertainties and private information protection of MUs. Firstly, a novel economic model for mobile crowdsensing is designed, which takes MUs' resources constraints and demand uncertainties into consideration. Secondly, an incentive mechanism based on Stackelberg game is provided, where the sensing-platform (SP) is the leader and the MUs are the followers. Then, the existence and uniqueness of the Stackelberg Equilibrium (SE) is proven and the procedure for computing the SE is given. Furthermore, a dynamic incentive mechanism (DIM) based on deep reinforcement learning (DRL) approach is investigated without knowing the private information of the MUs. It enables the SP to learn the optimal pricing strategy directly from game experience without any prior knowledge about MUs' information. Finally, numerical simulations are implemented to evaluate the performance and theoretical properties of the proposed mechanism and approach.

cs.CY↗

arXiv2018

A Survey of FPGA-Based Neural Network Accelerator

Recent researches on neural network have shown significant advantage in machine learning over traditional algorithms based on handcrafted features and models. Neural network is now widely adopted in regions like image, speech and video recognition. But the high computation and storage complexity of neural network inference poses great difficulty on its application. CPU platforms are hard to offer enough computation capacity. GPU platforms are the first choice for neural network process because of its high computation capacity and easy to use development frameworks. On the other hand, FPGA-based neural network inference accelerator is becoming a research topic. With specifically designed hardware, FPGA is the next possible solution to surpass GPU in speed and energy efficiency. Various FPGA-based accelerator designs have been proposed with software and hardware optimization techniques to achieve high speed and energy efficiency. In this paper, we give an overview of previous work on neural network inference accelerators based on FPGA and summarize the main techniques used. An investigation from software to hardware, from circuit level to system level is carried out to complete analysis of FPGA-based neural network inference accelerator design and serves as a guide to future work.

cs.AR↗

arXiv2018

Differentially Private User-based Collaborative Filtering Recommendation Based on K-means Clustering

Collaborative filtering (CF) recommendation algorithms are well-known for their outstanding recommendation performances, but previous researches showed that they could cause privacy leakage for users due to k-nearest neighboring (KNN) attacks. Recently, the notion of differential privacy (DP) has been applied to privacy preservation for collaborative filtering recommendation algorithms. However, as far as we know, existing differentially private CF recommendation schemes degraded the recommendation performance (such as recall and precision) to an unacceptable level. In this paper, in order to address the performance degradation problem, we propose a differentially private user-based collaborative filtering recommendation scheme based on k-means clustering (KDPCF). Specifically, to improve the recommendation performance, we first cluster the dataset into categories by k-means clustering and appropriately adjust the size of the target category to which the target user belongs, so that only users in the well-sized target category can be used for recommendations. Then we efficiently select a set of neighbors from the target category at one time by employing only one exponential mechanism instead of the composition of multiple ones, and base on the neighbor set to recommend. We theoretically prove that our scheme achieves differential privacy. Empirically, we use the MovieLens dataset to evaluate our recommendation system. The experimental results demonstrate a significant performance gain compared to existing schemes.

cs.CR↗

arXiv2018

Mg$_{3+δ}$Sb$_x$Bi$_{2-x}$ family: A promising substitute for the start-of-art n-type thermoelectric materials near room temperature

Bi2Te3-xSex family has been the n-type start-of-the-art thermoelectric materials near room temperatures (RT) for more than half-century, which dominates the active cooling and novel waves harvesting application near RT. However, the drawbacks of brittle nature and Te-containing restrict the further applications exploring. Here, we show that a Mg3+δSbxBi2-x family ((ZT)avg =1.05) could be a promising substitute for the Bi2Te3-xSex family ((ZT)avg =0.9-1.0) in the temperature range of 50-250 °C based on the comparable thermoelectric performance through a synergistic effect from the tunable band gap using the alloy effect and the suppressible Mg-vacancy formation using interstitial Mn dopant. The former is to shift the optimal thermoelectric performance to near RT, and latter is helpful to partially decouple the electrical transport and thermal transport in order to get an optimal RT power factor. A positive temperature-dependence of band gap suggested this family is also a superior medium-temperature thermoelectric material for the significantly suppressed bipolar effect. Furthermore, a two times higher mechanical toughness, compared with Bi2Te3-xSex family, consolidates the promising substitute for the start-of-art n-type thermoelectric materials near RT.

cond-mat.mtrl-sci↗

arXiv2018

Spin scattering and noncollinear spin structure-induced intrinsic anomalous Hall effect in antiferromagnetic topological insulator $\mathrm{MnBi_2Te_4}$

$\mathrm{MnBi_2Te_4}$ has recently been established as an intrinsic antiferromagnetic (AFM) topological insulator and predicted to be an ideal platform to realize quantum anomalous Hall (QAH) insulator and axion insulator states. We performed comprehensive studies on the structure, nontrivial surface state and magnetotransport properties of this material. Our results reveal an intrinsic anomalous Hall effect arising from a non-collinear spin structure for the magnetic field parallel to the $c$-axis. We also observed remarkable negative magnetoresistance under arbitrary field orientation below and above the Neel temperature (T$_N$), providing clear evidence for strong spin fluctuation-driven spin scattering in both the AFM and paramagnetic states. Further, we found that the nontrivial surface state opens a large gap (~85 meV) even at temperatures far above T$_N$ = 25K. These findings demonstrate that the bulk band structure of $\mathrm{MnBi_2Te_4}$ is strongly coupled to the magnetic structure and that a net Berry curvature in momentum space can be created in a canted AFM state. In addition, our results imply that the gap opening in the surface states is intrinsic, likely caused by the strong spin fluctuations near the surface layers.

cond-mat.mtrl-sci↗

arXiv2018

Experimental identification of critical condition for drastically enhancing thermoelectric power factor of two-dimensional layered materials

Nano-structuring is an extremely promising path to high performance thermoelectrics. Favorable improvements in thermal conductivity are attainable in many material systems, and theoretical work points to large improvements in electronic properties. However, realization of the electronic benefits in practical materials has been elusive experimentally. A key challenge is that experimental identification of the quantum confinement length, below which the thermoelectric power factor is significantly enhanced, remains elusive due to lack of simultaneous control of size and carrier density. Here we investigate gate tunable and temperature-dependent thermoelectric transport in $γ$ phase indium selenide ($γ$ InSe, n type semiconductor) samples with thickness varying from 7 to 29 nm. This allows us to properly map out dimension and doping space. Combining theoretical and experimental studies, we reveal that the sharper pre-edge of the conduction-band density of states arising from quantum confinement gives rise to an enhancement of the Seebeck coefficient and the power factor in the thinner InSe samples. Most importantly, we experimentally identify the role of the competition between quantum confinement length and thermal de Broglie wavelength in the enhancement of power factor. Our results provide an important and general experimental guideline for optimizing the power factor and improving the thermoelectric performance of two-dimensional layered semiconductors.

cond-mat.mes-hall↗

arXiv2018

Readout Electronics for CEPC Semi-Digital Hadron Calorimeter Pre-prototype

CEPC (Circular Electron and Positron Collider) is a large experiment facility proposed by Chinese particle physics community. One of its running option is being the Higgs factory. Calorimeter is the main part of this experiment to measure the jet energy. Semi-digital hadron calorimeter (SDHCAL) is one of the options for the hadron measurement. GEM detector with its high position resolution and flexible configuration is one of the candidates for the active layer of the SDHCAL. The main purpose of this paper is to provide a feasible readout method for the GEM-based semi-digital hadron calorimeter. A small-scale prototype is designed and implemented, including front-end board (FEB) and data interface board (DIF). The prototype electronics has been tested. The equivalent RMS noise of all channels is below 0.35fC. The dynamic range is up to 500fC and the gain variation is less than 1%. The readout electronics is applied on a double-layer GEM detector with 1cm*1cm readout pad. Result indicates that the electronics works well with the detector. The detection efficiency of MIP is over 95% with 5fC threshold.

physics.ins-det↗

arXiv2018

Proximity-Induced Superconductivity with Subgap Anomaly in Type II Weyl Semi-Metal WTe2

Due to the non-trivial topological band structure in type-II Weyl semimetal Tungsten ditelluride (WTe2), unconventional properties may emerge in its superconducting phase. While realizing intrinsic superconductivity has been challenging in the type-II Weyl semimetal WTe2, proximity effect may open an avenue for the realization of superconductivity. Here, we report the observation of proximity-induced superconductivity with a long coherence length along c axis in WTe2 thin flakes based on a WTe2/NbSe2 van der Waals heterostructure. Interestingly, we also observe anomalous oscillations of the differential resistance during the transition from superconducting to normal state. Theoretical calculations show excellent agreement with experimental results, revealing that such a sub-gap anomaly is the intrinsic property of WTe2 in superconducting state induced by the proximity effect. Our findings enrich the understanding of superconducting phase of type-II Weyl semimetals, and pave the way for their future applications in topological quantum computing.

cond-mat.supr-con↗

arXiv2018

Confidence Estimation and Deletion Prediction Using Bidirectional Recurrent Neural Networks

The standard approach to assess reliability of automatic speech transcriptions is through the use of confidence scores. If accurate, these scores provide a flexible mechanism to flag transcription errors for upstream and downstream applications. One challenging type of errors that recognisers make are deletions. These errors are not accounted for by the standard confidence estimation schemes and are hard to rectify in the upstream and downstream processing. High deletion rates are prominent in limited resource and highly mismatched training/testing conditions studied under IARPA Babel and Material programs. This paper looks at the use of bidirectional recurrent neural networks to yield confidence estimates in predicted as well as deleted words. Several simple schemes are examined for combination. To assess usefulness of this approach, the combined confidence score is examined for untranscribed data selection that favours transcriptions with lower deletion errors. Experiments are conducted using IARPA Babel/Material program languages.

eess.AS↗

arXiv2018

Interaction effects with varying $N$ in SU($N$) symmetric fermion lattice systems

The interaction effects in ultracold Fermi gases with SU($N$) symmetry are studied non-perturbatively in half-filled one-dimensional lattices by employing quantum Monte Carlo simulations.We find that as $N$ increases, weak and strong interacting systems are driven to a crossover region, but from opposite directions as a convergence of itinerancy and Mottness.In the weak interaction region, particles are nearly itinerant,and inter-particle collisions are enhanced by $N$, resulting in the amplification of interaction effects. In contrast, in the strong coupling region, increasing $N$ softens the Mott-insulating background through the enhanced virtual hopping processes.The crossover region exhibits nearly $N$-independent physical quantities, including the relative bandwidth, Fermi distribution, and the spin structure factor.The difference between even-$N$ and odd-$N$ systems is most prominent at small $N$'s with strong interactions, since the odd case allows local real hopping with an energy scale much larger than the virtual one.The above effects can be experimentally tested in ultracold atom experiments with alkaline-earth (-like) fermions such as $^{87}$Sr ($^{173}$Yb).

cond-mat.str-el↗

arXiv2018

A Neural Transition-based Model for Nested Mention Recognition

It is common that entity mentions can contain other mentions recursively. This paper introduces a scalable transition-based method to model the nested structure of mentions. We first map a sentence with nested mentions to a designated forest where each mention corresponds to a constituent of the forest. Our shift-reduce based system then learns to construct the forest structure in a bottom-up manner through an action sequence whose maximal length is guaranteed to be three times of the sentence length. Based on Stack-LSTM which is employed to efficiently and effectively represent the states of the system in a continuous space, our system is further incorporated with a character-based component to capture letter-level patterns. Our model achieves the state-of-the-art results on ACE datasets, showing its effectiveness in detecting nested mentions.

cs.CL↗

arXiv2018

Electron mass enhancement and magnetic phase separation near the Mott transition in double layer ruthenates

We present a detailed investigation of the specific heat in Ca$_3$(Ru$_{1-x}M_x$)$_2$O$_7$ ($M$ = Ti, Fe, Mn) single crystals. With different dopants and doping levels, three distinct regions are present, including a quasi-2D metallic state with an antiferromagnetic (AFM) order formed by ferromagnetic bilayers (AFM-$b$), a Mott insulating state with G-type AFM order (G-AFM) and a localized state with a mixed AFM-b and G-AFM phase. Our specific heat data provide deep insights into the Mott transitions induced by Ti and Mn dopings. We observed not only an anomalous large mass enhancement but also an additional term in the specific heat i.e. $C\propto T^2$ in the localized region. The $C\propto T^2$ term is most likely due to the long-wavelength excitations with both FM and AFM components. A decrease of Debye temperature is observed in the G-type AFM region, indicating a lattice softening associated with the Mott transition.

cond-mat.str-el↗

arXiv2018

Direct observation of ferroelectricity in Ca$_3$Mn$_2$O$_7$ and its prominent light absorption

Layered perovskites $A_3M_2$O$_7$ are known to exhibit the so-called hybrid improper ferroelectricity. Despite experimentally confirmed cases (e.g. nonmagnetic $M$=Ti and Sn), the ferroelectricity in magnetic Ca$_3$Mn$_2$O$_7$ remains a puzzle. Here, the structural, ferroelectric, magnetoelectric, and optical properties of Ca$_3$Mn$_2$O$_7$ are systematically investigated. Switchable polarization is directly measured, demonstrating its ferroelectricity. In addition, magnetoelectric response is also evidenced, implying the coupling between magnetism and ferroelectricity. Furthermore, strong visible light absorption is observed, which can be understood from its electronic structure. Its direct and appropriate band gap, as well as wide conducting bands, makes Ca$_3$Mn$_2$O$_7$ a potential candidate for ferroelectric photoelectric applications.

cond-mat.mtrl-sci↗

arXiv2018

User Information Augmented Semantic Frame Parsing using Coarse-to-Fine Neural Networks

Semantic frame parsing is a crucial component in spoken language understanding (SLU) to build spoken dialog systems. It has two main tasks: intent detection and slot filling. Although state-of-the-art approaches showed good results, they require large annotated training data and long training time. In this paper, we aim to alleviate these drawbacks for semantic frame parsing by utilizing the ubiquitous user information. We design a novel coarse-to-fine deep neural network model to incorporate prior knowledge of user information intermediately to better and quickly train a semantic frame parser. Due to the lack of benchmark dataset with real user information, we synthesize the simplest type of user information (location and time) on ATIS benchmark data. The results show that our approach leverages such simple user information to outperform state-of-the-art approaches by 0.25% for intent detection and 0.31% for slot filling using standard training data. When using smaller training data, the performance improvement on intent detection and slot filling reaches up to 1.35% and 1.20% respectively. We also show that our approach can achieve similar performance as state-of-the-art approaches by using less than 80% annotated training data. Moreover, the training time to achieve the similar performance is also reduced by over 60%.

cs.CL↗