scispace - formally typeset
Search or ask a question

Showing papers by "Beihang University published in 2017"


Journal ArticleDOI
Zheng Zhao1, Weihai Chen1, Xingming Wu1, Peter C. Y. Chen, Jingmeng Liu1 
TL;DR: A novel traffic forecast model based on long short-term memory (LSTM) network is proposed, which considers temporal-spatial correlation in traffic system via a two-dimensional network which is composed of many memory units.
Abstract: Short-term traffic forecast is one of the essential issues in intelligent transportation system. Accurate forecast result enables commuters make appropriate travel modes, travel routes, and departure time, which is meaningful in traffic management. To promote the forecast accuracy, a feasible way is to develop a more effective approach for traffic data analysis. The availability of abundant traffic data and computation power emerge in recent years, which motivates us to improve the accuracy of short-term traffic forecast via deep learning approaches. A novel traffic forecast model based on long short-term memory (LSTM) network is proposed. Different from conventional forecast models, the proposed LSTM network considers temporal-spatial correlation in traffic system via a two-dimensional network which is composed of many memory units. A comparison with other representative forecast models validates that the proposed LSTM network can achieve a better performance.

1,204 citations


Journal ArticleDOI
TL;DR: In this paper, the historical development, new phenomena and emerging applications of superwettability systems are discussed and a review of the superwetability properties of interfacial materials is presented.
Abstract: Studying nature to reveal the mechanisms of special wetting phenomena in biological systems can effectively inspire the design and fabrication of functional interfacial materials with superwettability. In this Review, the historical development, new phenomena and emerging applications of superwettability systems are discussed.

1,109 citations


Journal ArticleDOI
TL;DR: It is demonstrated that the Li2S decomposition energy barrier is associated with the binding between isolated Li ions and the sulfur in sulfides; this is the main reason that sulfide materials can induce lower overpotential compared with commonly used carbon materials.
Abstract: Polysulfide binding and trapping to prevent dissolution into the electrolyte by a variety of materials has been well studied in Li−S batteries. Here we discover that some of those materials can play an important role as an activation catalyst to facilitate oxidation of the discharge product, Li2S, back to the charge product, sulfur. Combining theoretical calculations and experimental design, we select a series of metal sulfides as a model system to identify the key parameters in determining the energy barrier for Li2S oxidation and polysulfide adsorption. We demonstrate that the Li2S decomposition energy barrier is associated with the binding between isolated Li ions and the sulfur in sulfides; this is the main reason that sulfide materials can induce lower overpotential compared with commonly used carbon materials. Fundamental understanding of this reaction process is a crucial step toward rational design and screening of materials to achieve high reversible capacity and long cycle life in Li−S batteries.

933 citations


Journal ArticleDOI
TL;DR: Using tetrahexahedral gold nanorods as a heterogeneous electrocatalyst, an electrocatalytic N2 reduction reaction was shown to be possible at room temperature and atmospheric pressure, with a high Faradic efficiency up to 4.02% at -0.2 V vs reversible hydrogen electrode.
Abstract: Using tetrahexahedral gold nanorods as a heterogeneous electrocatalyst, an electrocatalytic N2 reduction reaction is shown to be possible at room temperature and atmospheric pressure, with a high Faradic efficiency up to 4.02% at -0.2 V vs reversible hydrogen electrode (1.648 µg h-1 cm-2 and 0.102 µg h-1 cm-2 for NH3 and N2 H4 ·H2 O, respectively).

923 citations


Journal ArticleDOI
10 Apr 2017-Sensors
TL;DR: Wang et al. as mentioned in this paper proposed a convolutional neural network (CNN)-based method that learns traffic as images and predicts large-scale, network-wide traffic speed with a high accuracy.
Abstract: This paper proposes a convolutional neural network (CNN)-based method that learns traffic as images and predicts large-scale, network-wide traffic speed with a high accuracy. Spatiotemporal traffic dynamics are converted to images describing the time and space relations of traffic flow via a two-dimensional time-space matrix. A CNN is applied to the image following two consecutive steps: abstract traffic feature extraction and network-wide traffic speed prediction. The effectiveness of the proposed method is evaluated by taking two real-world transportation networks, the second ring road and north-east transportation network in Beijing, as examples, and comparing the method with four prevailing algorithms, namely, ordinary least squares, k-nearest neighbors, artificial neural network, and random forest, and three deep learning architectures, namely, stacked autoencoder, recurrent neural network, and long-short-term memory network. The results show that the proposed method outperforms other algorithms by an average accuracy improvement of 42.91% within an acceptable execution time. The CNN can train the model in a reasonable time and, thus, is suitable for large-scale transportation networks.

894 citations


Journal ArticleDOI
TL;DR: Owing to the interfacial interaction between BP and CN, efficient charge transfer occurred, thereby enhancing the photocatalytic performance, and the present results show that BP/CN is a metal-free photocatalyst for artificial photosynthesis and renewable energy conversion.
Abstract: In the drive toward green and sustainable chemistry, exploring efficient and stable metal-free photocatalysts with broadband solar absorption from the UV to near-infrared region for the photoreduction of water to H2 remains a big challenge. To this end, a binary nanohybrid (BP/CN) of two-dimensional (2D) black phosphorus (BP) and graphitic carbon nitride (CN) was designed and used as a metal-free photocatalyst for the first time. During irradiation of BP/CN in water with >420 and >780 nm light, solid H2 gas was generated, respectively. Owing to the interfacial interaction between BP and CN, efficient charge transfer occurred, thereby enhancing the photocatalytic performance. The efficient charge-trapping and transfer processes were thoroughly investigated with time-resolved diffuse reflectance spectroscopic measurement. The present results show that BP/CN is a metal-free photocatalyst for artificial photosynthesis and renewable energy conversion.

857 citations


Journal ArticleDOI
Fei Tao1, Meng Zhang1
TL;DR: A novel concept of digital twin shop-floor (DTS) based on digital twin is explored and its four key components are discussed, including physicalShop-floor, virtual shop- Floor, shop- floor service system, and shop-ground digital twin data.
Abstract: With the developments and applications of the new information technologies, such as cloud computing, Internet of Things, big data, and artificial intelligence, a smart manufacturing era is coming. At the same time, various national manufacturing development strategies have been put forward, such as Industry 4.0 , Industrial Internet , manufacturing based on Cyber-Physical System , and Made in China 2025 . However, one of specific challenges to achieve smart manufacturing with these strategies is how to converge the manufacturing physical world and the virtual world, so as to realize a series of smart operations in the manufacturing process, including smart interconnection, smart interaction, smart control and management, etc. In this context, as a basic unit of manufacturing, shop-floor is required to reach the interaction and convergence between physical and virtual spaces, which is not only the imperative demand of smart manufacturing, but also the evolving trend of itself. Accordingly, a novel concept of digital twin shop-floor (DTS) based on digital twin is explored and its four key components are discussed, including physical shop-floor, virtual shop-floor, shop-floor service system, and shop-floor digital twin data. What is more, the operation mechanisms and implementing methods for DTS are studied and key technologies as well as challenges ahead are investigated, respectively.

741 citations


Journal ArticleDOI
Albert M. Sirunyan, Armen Tumasyan, Wolfgang Adam1, Ece Aşılar1  +2212 moreInstitutions (157)
TL;DR: A fully-fledged particle-flow reconstruction algorithm tuned to the CMS detector was developed and has been consistently used in physics analyses for the first time at a hadron collider as mentioned in this paper.
Abstract: The CMS apparatus was identified, a few years before the start of the LHC operation at CERN, to feature properties well suited to particle-flow (PF) reconstruction: a highly-segmented tracker, a fine-grained electromagnetic calorimeter, a hermetic hadron calorimeter, a strong magnetic field, and an excellent muon spectrometer. A fully-fledged PF reconstruction algorithm tuned to the CMS detector was therefore developed and has been consistently used in physics analyses for the first time at a hadron collider. For each collision, the comprehensive list of final-state particles identified and reconstructed by the algorithm provides a global event description that leads to unprecedented CMS performance for jet and hadronic τ decay reconstruction, missing transverse momentum determination, and electron and muon identification. This approach also allows particles from pileup interactions to be identified and enables efficient pileup mitigation methods. The data collected by CMS at a centre-of-mass energy of 8\TeV show excellent agreement with the simulation and confirm the superior PF performance at least up to an average of 20 pileup interactions.

719 citations


Journal ArticleDOI
TL;DR: An effective and reliable deep learning method known as stacked denoising autoencoder (SDA), which is shown to be suitable for certain health state identifications for signals containing ambient noise and working condition fluctuations, is investigated.

591 citations


Journal ArticleDOI
TL;DR: Time-varying formation tracking analysis and design problems for second-order Multi-Agent systems with switching interaction topologies are studied, and a formation tracking protocol is constructed based on the relative information of the neighboring agents.
Abstract: Time-varying formation tracking analysis and design problems for second-order Multi-Agent systems with switching interaction topologies are studied, where the states of the followers form a predefined time-varying formation while tracking the state of the leader. A formation tracking protocol is constructed based on the relative information of the neighboring agents. Necessary and sufficient conditions for Multi-Agent systems with switching interaction topologies to achieve time-varying formation tracking are proposed together with the formation tracking feasibility constraint based on the graph theory. An approach to design the formation tracking protocol is proposed by solving an algebraic Riccati equation, and the stability of the proposed approach is proved using the common Lyapunov stability theory. The obtained results are applied to solve the target enclosing problem of a multiquadrotor unmanned aerial vehicle (UAV) system consisting of one leader (target) quadrotor UAV and three follower quadrotor UAVs. A numerical simulation and an outdoor experiment are presented to demonstrate the effectiveness of the theoretical results.

566 citations


Proceedings ArticleDOI
29 Jul 2017
TL;DR: The authors proposed to inject corpus-level constraints for calibrating existing structured prediction models and design an algorithm based on Lagrangian relaxation for collective inference, which results in almost no performance loss for the underlying recognition task but decreases the magnitude of bias amplification.
Abstract: Language is increasingly being used to de-fine rich visual recognition problems with supporting image collections sourced from the web. Structured prediction models are used in these tasks to take advantage of correlations between co-occurring labels and visual input but risk inadvertently encoding social biases found in web corpora. In this work, we study data and models associated with multilabel object classification and visual semantic role labeling. We find that (a) datasets for these tasks contain significant gender bias and (b) models trained on these datasets further amplify existing bias. For example, the activity cooking is over 33% more likely to involve females than males in a training set, and a trained model further amplifies the disparity to 68% at test time. We propose to inject corpus-level constraints for calibrating existing structured prediction models and design an algorithm based on Lagrangian relaxation for collective inference. Our method results in almost no performance loss for the underlying recognition task but decreases the magnitude of bias amplification by 47.5% and 40.5% for multilabel classification and visual semantic role labeling, respectively。

Journal ArticleDOI
TL;DR: In this article, the authors discuss the latest advancements in the 2D layered inorganic materials for gas sensors and introduce various types of gas-sensing devices and describe the basic parameters and influence factors of the gas sensors to further enhance their performance.
Abstract: Two-dimensional (2D) layered inorganic nanomaterials have attracted huge attention due to their unique electronic structures, as well as extraordinary physical and chemical properties for use in electronics, optoelectronics, spintronics, catalysts, energy generation and storage, and chemical sensors Graphene and related layered inorganic analogues have shown great potential for gas-sensing applications because of their large specific surface areas and strong surface activities This review aims to discuss the latest advancements in the 2D layered inorganic materials for gas sensors We first elaborate the gas-sensing mechanisms and introduce various types of gas-sensing devices Then, we describe the basic parameters and influence factors of the gas sensors to further enhance their performance Moreover, we systematically present the current gas-sensing applications based on graphene, graphene oxide (GO), reduced graphene oxide (rGO), functionalized GO or rGO, transition metal dichalcogenides, layered II

Journal ArticleDOI
Qinfeng Rong1, Wenwei Lei1, Lie Chen1, Yong-ai Yin1, Jiajia Zhou1, Mingjie Liu1 
TL;DR: Anti-freezing conductive organohydrogels are reported by using an H2 O/ethylene glycol binary solvent as dispersion medium with non-covalent crosslinks to exhibit stable flexibility and strain-sensitivity in the temperature range from -55.0 to 44.6 °C.
Abstract: Conductive hydrogels are a class of stretchable conductive materials that are important for various applications. However, water-based conductive hydrogels inevitably lose elasticity and conductivity at subzero temperatures, which severely limits their applications at low temperatures. Herein we report anti-freezing conductive organohydrogels by using an H2O/ethylene glycol binary solvent as dispersion medium. Owing to the freezing tolerance of the binary solvent, our organohydrogels exhibit stable flexibility and strain-sensitivity in the temperature range from −55.0 to 44.6 °C. Meanwhile, the solvent molecules could form hydrogen bonds with polyvinyl alcohol (PVA) chains and induce the crystallization of PVA, greatly improving the mechanical strength of the organohydrogels. Furthermore, the non-covalent crosslinks endow the conductive organohydrogels with intriguing remoldability and self-healing capability, which are important for practical applications.

Proceedings ArticleDOI
Yu Wu1, Wei Wu2, Chen Xing3, Ming Zhou2, Zhoujun Li1 
05 Aug 2017
TL;DR: Zhang et al. as mentioned in this paper proposed a sequential matching network (SMN) which first matches a response with each utterance in the context on multiple levels of granularity, and distills important matching information from each pair as a vector with convolution and pooling operations.
Abstract: We study response selection for multi-turn conversation in retrieval based chatbots. Existing work either concatenates utterances in context or matches a response with a highly abstract context vector finally, which may lose relationships among the utterances or important information in the context. We propose a sequential matching network (SMN) to address both problems. SMN first matches a response with each utterance in the context on multiple levels of granularity, and distills important matching information from each pair as a vector with convolution and pooling operations. The vectors are then accumulated in a chronological order through a recurrent neural network (RNN) which models relationships among the utterances. The final matching score is calculated with the hidden states of the RNN. Empirical study on two public data sets shows that SMN can significantly outperform state-of-the-art methods for response selection in multi-turn conversation.

Journal ArticleDOI
TL;DR: Experimental results indicate that the effective integration of high catalytic reactivity, high structural stability, and high electronic conductivity into a single material system makes Ni-Fe-OH@Ni3 S2 /NF a remarkable catalytic ability for OER at large current densities.
Abstract: Developing nonprecious oxygen evolution electrocatalysts that can work well at large current densities is of primary importance in a viable water-splitting technology. Herein, a facile ultrafast (5 s) synthetic approach is reported that produces a novel, efficient, non-noble metal oxygen-evolution nano-electrocatalyst that is composed of amorphous Ni-Fe bimetallic hydroxide film-coated, nickel foam (NF)-supported, Ni3 S2 nanosheet arrays. The composite nanomaterial (denoted as Ni-Fe-OH@Ni3 S2 /NF) shows highly efficient electrocatalytic activity toward oxygen evolution reaction (OER) at large current densities, even in the order of 1000 mA cm-2 . Ni-Fe-OH@Ni3 S2 /NF also gives an excellent catalytic stability toward OER both in 1 m KOH solution and in 30 wt% KOH solution. Further experimental results indicate that the effective integration of high catalytic reactivity, high structural stability, and high electronic conductivity into a single material system makes Ni-Fe-OH@Ni3 S2 /NF a remarkable catalytic ability for OER at large current densities.

Journal ArticleDOI
TL;DR: When dealing with uncertainties, it is shown that DUEA has a different but complementary mechanism to widely used robust control and adaptive control and other promising methods such as internal model control and output regulation theory.
Abstract: This paper gives a comprehensive overview on disturbance/uncertainty estimation and attenuation (DUEA) techniques in permanent-magnet synchronous motor (PMSM) drives. Various disturbances and uncertainties in PMSM and also other alternating current (ac) motor drives are first reviewed which shows they have different behaviors and appear in different control loops of the system. The existing DUEA and other relevant control methods in handling disturbances and uncertainties widely used in PMSM drives, and their latest developments are then discussed and summarized. It also provides in-depth analysis of the relationship between these advanced control methods in the context of PMSM systems. When dealing with uncertainties, it is shown that DUEA has a different but complementary mechanism to widely used robust control and adaptive control. The similarities and differences in disturbance attenuation of DUEA and other promising methods such as internal model control and output regulation theory have been analyzed in detail. The wide applications of these methods in different ac motor drives (in particular in PMSM drives) are categorized and summarized. Finally, the paper ends with the discussion on future directions in this area.

Journal ArticleDOI
Haiyang Yu1, Wu Zhihai1, Shuqin Wang, Yunpeng Wang1, Xiaolei Ma1 
26 Jun 2017-Sensors
TL;DR: Wang et al. as mentioned in this paper proposed a spatiotemporal recurrent convolutional networks (SRCNs) for traffic forecasting, which inherit the advantages of deep CNNs and LSTM neural networks.
Abstract: Predicting large-scale transportation network traffic has become an important and challenging topic in recent decades. Inspired by the domain knowledge of motion prediction, in which the future motion of an object can be predicted based on previous scenes, we propose a network grid representation method that can retain the fine-scale structure of a transportation network. Network-wide traffic speeds are converted into a series of static images and input into a novel deep architecture, namely, spatiotemporal recurrent convolutional networks (SRCNs), for traffic forecasting. The proposed SRCNs inherit the advantages of deep convolutional neural networks (DCNNs) and long short-term memory (LSTM) neural networks. The spatial dependencies of network-wide traffic can be captured by DCNNs, and the temporal dynamics can be learned by LSTMs. An experiment on a Beijing transportation network with 278 links demonstrates that SRCNs outperform other deep learning-based algorithms in both short-term and long-term traffic prediction.

Journal ArticleDOI
TL;DR: Numerical simulations are provided to demonstrate the effectiveness of the theoretical results and the obtained results can be applied to deal with time-varying formation tracking problems, target enclosing problems, and consensus tracking problems for linear multiagent systems with one or multiple targets/leaders.
Abstract: Time-varying formation tracking problems for linear multiagent systems with multiple leaders are studied, where the states of followers form a predefined time-varying formation while tracking the convex combination of the states of multiple leaders. Followers are classified into well-informed ones and uninformed ones, where the neighbor set of the former contains all the leaders, whereas the latter contains no leaders. A formation tracking protocol is constructed using only neighboring relative information. Necessary and sufficient conditions for multiagent systems with multiple leaders to achieve time-varying formation tracking are proposed by utilizing the properties of the Laplacian matrix, where the formation tracking feasibility constraints are also given. An approach to design the formation tracking protocol is presented by solving an algebraic Riccati equation. The obtained results can be applied to deal with time-varying formation tracking problems, target enclosing problems, and consensus tracking problems for linear multiagent systems with one or multiple targets/leaders. Numerical simulations are provided to demonstrate the effectiveness of the theoretical results.

Journal ArticleDOI
TL;DR: Five polymer donors with distinct chemical structures and different electronic properties are surveyed in a planar and narrow-bandgap fused-ring electron acceptor (IDIC)-based organic solar cells, which exhibit power conversion efficiencies of up to 11%.
Abstract: Five polymer donors with distinct chemical structures and different electronic properties are surveyed in a planar and narrow-bandgap fused-ring electron acceptor (IDIC)-based organic solar cells, which exhibit power conversion efficiencies of up to 11%.

Journal ArticleDOI
S. Hirose1, T. Iijima1, I. Adachi2, K. Adamczyk  +190 moreInstitutions (61)
TL;DR: The first measurement of the tau lepton polarization P-tau(D*) in the decay (B) over bar -> D* tau(-) (v) over b (tau) as well as a new measurement of the ratio of the branching fractions was reported in this paper.
Abstract: We report the first measurement of the tau lepton polarization P-tau(D*) in the decay (B) over bar -> D* tau(-) (v) over bar (tau) as well as a newmeasurement of the ratio of the branching fractions R(D*) = B((B) over bar -> D* tau(-) (v) over bar (tau)) / B((B) over bar -> D* l(-) (v) over bar (l)), where l(-) denotes an electron or a muon, and the tau is reconstructed in the modes tau(-) -> pi(-) v(tau) and tau(-) -> rho(-) v(tau). We use the full data sample of 772 x 10(6) B (B) over bar pairs recorded with the Belle detector at the (KEKB) over bar electron-positron collider. Our results, P-tau(D*) = -0.38 +/- 0.51 (stat)(-0.16)(+0.21) (syst) and R(D*) = 0.270 +/- 0.035 (stat)(- 0.025)(+0.028) (syst), are consistent with the theoretical predictions of the standard model.

Journal ArticleDOI
TL;DR: Superefficient water-splitting materials comprising sub-nanometric copper clusters and quasi-amorphous cobalt sulfide supported on copper foam give a catalytic output of overall water splitting comparable with the Pt/C-IrO2 -coupled electrolyzer.
Abstract: Superefficient water-splitting materials comprising sub-nanometric copper clusters and quasi-amorphous cobalt sulfide supported on copper foam are reported. While working together at both the anode and cathode sides of an alkaline electrolyzer, this material gives a catalytic output of overall water splitting comparable with the Pt/C-IrO2 -coupled electrolyzer.

Journal ArticleDOI
TL;DR: A facile approach is reported to transform wood into hierarchical porous graphene using CO2 laser scribing to inspire both research and industrial interest in the development of wood-derived graphene materials and their nanodevices.
Abstract: Wood as a renewable naturally occurring resource has been the focus of much research and commercial interests in applications ranging from building construction to chemicals production. Here, a facile approach is reported to transform wood into hierarchical porous graphene using CO2 laser scribing. Studies reveal that the crosslinked lignocellulose structure inherent in wood with higher lignin content is more favorable for the generation of high-quality graphene than wood with lower lignin content. Because of its high electrical conductivity (≈10 Ω per square), graphene patterned on wood surfaces can be readily fabricated into various high-performance devices, such as hydrogen evolution and oxygen evolution electrodes for overall water splitting with high reaction rates at low overpotentials, and supercapacitors for energy storage with high capacitance. The versatility of this technique in formation of multifunctional wood hybrids can inspire both research and industrial interest in the development of wood-derived graphene materials and their nanodevices.

Journal ArticleDOI
TL;DR: In this paper, the ozone removal rate of α-MnO2 nanofiber with high concentration of surface oxygen vacancy was obtained via vacuum deoxidation method, where the formation of oxygen vacancy enhanced the ratio of Mn3+/Mn4+, resulting in a significant improvement of the adsorption of ozone on the surface of the catalyst.
Abstract: α-MnO2 nanofiber with high concentration of surface oxygen vacancy was obtained via vacuum deoxidation method. The activity of α-MnO2 strongly depends on the concentration and extent of oxygen vacancy, which can be adjusted by tuning the temperature and time of vacuum deoxidation. The formation of oxygen vacancy enhanced the ratio of Mn3+/Mn4+, which changed the charge distribution on the α-MnO2 nanofiber, resulting in a significant improvement of the adsorption of ozone on the surface of the catalyst. In the dry gas flow, the ozone removal rate at 20 h has increased from 32.6% to 95%. In the wet gas flow, the ozone removal rate was also enhanced thanks to more active sites offered by α-MnO2. What's more, we found that the deactivation caused by water vapor was temporary and the activity would recover once the humidity has decreased. Finally, DFT calculation revealed that surface oxygen vacancy was the adsorption and reaction site for ozone decomposition and a new mechanism of ozone decomposition in the presence of H2O also was proposed. This work developed a deeper understanding to the process of ozone decomposition and would promote manganese oxide catalyst for practical application.

Proceedings Article
12 Feb 2017
TL;DR: The authors proposed a topic aware sequence-to-sequence (TA-Seq2Seq) model, which utilizes topics to simulate prior human knowledge that guides them to form informative and interesting responses in conversation.
Abstract: We consider incorporating topic information into a sequence-to-sequence framework to generate informative and interesting responses for chatbots. To this end, we propose a topic aware sequence-to-sequence (TA-Seq2Seq) model. The model utilizes topics to simulate prior human knowledge that guides them to form informative and interesting responses in conversation, and leverages topic information in generation by a joint attention mechanism and a biased generation probability. The joint attention mechanism summarizes the hidden vectors of an input message as context vectors by message attention and synthesizes topic vectors by topic attention from the topic words of the message obtained from a pre-trained LDA model, with these vectors jointly affecting the generation of words in decoding. To increase the possibility of topic words appearing in responses, the model modifies the generation probability of topic words by adding an extra probability item to bias the overall distribution. Empirical studies on both automatic evaluation metrics and human annotations show that TA-Seq2Seq can generate more informative and interesting responses, significantly outperforming state-of-the-art response generation models.

Journal ArticleDOI
TL;DR: A model of the influence of customer engagement on stickiness is presented, showing that customer engagement has a direct and positive influence on customer stickiness as well as an indirect influence through customer value creation.

Posted Content
Haiyang Yu1, Wu Zhihai1, Shuqin Wang, Yunpeng Wang1, Xiaolei Ma1 
TL;DR: A network grid representation method that can retain the fine-scale structure of a transportation network and outperform other deep learning-based algorithms in both short-term and long-term traffic prediction is proposed.
Abstract: Predicting large-scale transportation network traffic has become an important and challenging topic in recent decades. Inspired by the domain knowledge of motion prediction, in which the future motion of an object can be predicted based on previous scenes, we propose a network grid representation method that can retain the fine-scale structure of a transportation network. Network-wide traffic speeds are converted into a series of static images and input into a novel deep architecture, namely, spatiotemporal recurrent convolutional networks (SRCNs), for traffic forecasting. The proposed SRCNs inherit the advantages of deep convolutional neural networks (DCNNs) and long short-term memory (LSTM) neural networks. The spatial dependencies of network-wide traffic can be captured by DCNNs, and the temporal dynamics can be learned by LSTMs. An experiment on a Beijing transportation network with 278 links demonstrates that SRCNs outperform other deep learning-based algorithms in both short-term and long-term traffic prediction.

Journal ArticleDOI
S. Wehle, C. Niebuhr, S. Yashchenko, Iki Adachi1  +239 moreInstitutions (64)
TL;DR: The result is consistent with standard model (SM) expectations, where the largest discrepancy from a SM prediction is observed in the muon modes with a local significance of 2.6σ.
Abstract: We present a measurement of angular observables and a test of lepton flavor universality in the B -> K(+)l(+)l(-) decay, where l is either e or mu. The analysis is performed on a data sample corresponding to an integrated luminosity of 711 fb(-1) containing 772 x 10(6) B (B) over bar pairs, collected at the Upsilon(4S) resonance with the Belle detector at the asymmetric-energy e(+)e(-) collider KEKB. The result is consistent with standard model (SM) expectations, where the largest discrepancy from a SM prediction is observed in the muon modes with a local significance of 2.6 sigma.

Proceedings ArticleDOI
01 Jul 2017
TL;DR: This paper introduces a dataset, denoted as MAFA, with 30, 811 Internet images and 35, 806 masked faces, and proposes LLE-CNNs for masked face detection, which consist of three major modules.
Abstract: Detecting faces with occlusions is a challenging task due to two main reasons: 1) the absence of large datasets of masked faces, and 2) the absence of facial cues from the masked regions. To address these two issues, this paper first introduces a dataset, denoted as MAFA, with 30, 811 Internet images and 35, 806 masked faces. Faces in the dataset have various orientations and occlusion degrees, while at least one part of each face is occluded by mask. Based on this dataset, we further propose LLE-CNNs for masked face detection, which consist of three major modules. The Proposal module first combines two pre-trained CNNs to extract candidate facial regions from the input image and represent them with high dimensional descriptors. After that, the Embedding module is incorporated to turn such descriptors into a similarity-based descriptor by using locally linear embedding (LLE) algorithm and the dictionaries trained on a large pool of synthesized normal faces, masked faces and non-faces. In this manner, many missing facial cues can be largely recovered and the influences of noisy cues introduced by diversified masks can be greatly alleviated. Finally, the Verification module is incorporated to identify candidate facial regions and refine their positions by jointly performing the classification and regression tasks within a unified CNN. Experimental results on the MAFA dataset show that the proposed approach remarkably outperforms 6 state-of-the-arts by at least 15.6%.

Journal ArticleDOI
TL;DR: A novel strategy to directly bioprint cell‐laden gelatin methacryloyl (GelMA) constructs using bioinks of GelMA physical gels (GPGs) achieved through a simple cooling process is reported.
Abstract: Bioprinting is an emerging technique for the fabrication of 3D cell-laden constructs. However, the progress for generating a 3D complex physiological microenvironment has been hampered by a lack of advanced cell-responsive bioinks that enable bioprinting with high structural fidelity, particularly in the case of extrusion-based bioprinting. Herein, this paper reports a novel strategy to directly bioprint cell-laden gelatin methacryloyl (GelMA) constructs using bioinks of GelMA physical gels (GPGs) achieved through a simple cooling process. Attributed to their shear-thinning and self-healing properties, the GPG bioinks can retain the shape and form integral structures after deposition, allowing for subsequent UV crosslinking for permanent stabilization. This paper shows the structural fidelity by bioprinting various 3D structures that are typically challenging to fabricate using conventional bioinks under extrusion modes. Moreover, the use of the GPG bioinks enables direct bioprinting of highly porous and soft constructs at relatively low concentrations (down to 3%) of GelMA. It is also demonstrated that the bioprinted constructs not only permit cell survival but also enhance cell proliferation as well as spreading at lower concentrations of the GPG bioinks. It is believed that such a strategy of bioprinting will provide many opportunities in convenient fabrication of 3D cell-laden constructs for applications in tissue engineering, regenerative medicine, and pharmaceutical screening.

Journal ArticleDOI
TL;DR: By engineering strain into cobalt oxide, the authors transform a once poor hydrogen evolution catalyst into one that is competitive with the state of the art.
Abstract: Designing high-performance and cost-effective electrocatalysts toward oxygen evolution and hydrogen evolution reactions in water-alkali electrolyzers is pivotal for large-scale and sustainable hydrogen production. Earth-abundant transition metal oxide-based catalysts are particularly active for oxygen evolution reaction; however, they are generally considered inactive toward hydrogen evolution reaction. Here, we show that strain engineering of the outermost surface of cobalt(II) oxide nanorods can turn them into efficient electrocatalysts for the hydrogen evolution reaction. They are competitive with the best electrocatalysts for this reaction in alkaline media so far. Our theoretical and experimental results demonstrate that the tensile strain strongly couples the atomic, electronic structure properties and the activity of the cobalt(II) oxide surface, which results in the creation of a large quantity of oxygen vacancies that facilitate water dissociation, and fine tunes the electronic structure to weaken hydrogen adsorption toward the optimum region.