• Semi-supervised learning is attracting increasing attention due to the fact that datasets of many domains lack enough labeled data. Variational Auto-Encoder (VAE), in particular, has demonstrated the benefits of semi-supervised learning. The majority of existing semi-supervised VAEs utilize a classifier to exploit label information, where the parameters of the classifier are introduced to the VAE. Given the limited labeled data, learning the parameters for the classifiers may not be an optimal solution for exploiting label information. Therefore, in this paper, we develop a novel approach for semi-supervised VAE without classifier. Specifically, we propose a new model called Semi-supervised Disentangled VAE (SDVAE), which encodes the input data into disentangled representation and non-interpretable representation, then the category information is directly utilized to regularize the disentangled representation via the equality constraint. To further enhance the feature learning ability of the proposed VAE, we incorporate reinforcement learning to relieve the lack of data. The dynamic framework is capable of dealing with both image and text data with its corresponding encoder and decoder networks. Extensive experiments on image and text datasets demonstrate the effectiveness of the proposed framework.
  • The automatic parking is being massively developed by car manufacturers and providers. Until now, there are two problems with the automatic parking. First, there is no openly-available segmentation labels of parking slot on panoramic surround view (PSV) dataset. Second, how to detect parking slot and road structure robustly. Therefore, in this paper, we build up a public PSV dataset. At the same time, we proposed a highly fused convolutional network (HFCN) based segmentation method for parking slot and lane markings based on the PSV dataset. A surround-view image is made of four calibrated images captured from four fisheye cameras. We collect and label more than 4,200 surround view images for this task, which contain various illuminated scenes of different types of parking slots. A VH-HFCN network is proposed, which adopts an HFCN as the base, with an extra efficient VH-stage for better segmenting various markings. The VH-stage consists of two independent linear convolution paths with vertical and horizontal convolution kernels respectively. This modification enables the network to robustly and precisely extract linear features. We evaluated our model on the PSV dataset and the results showed outstanding performance in ground markings segmentation. Based on the segmented markings, parking slots and lanes are acquired by skeletonization, hough line transform and line arrangement.
  • TiEV is an autonomous driving platform implemented by Tongji University of China. The vehicle is drive-by-wire and is fully powered by electricity. We devised the software system of TiEV from scratch, which is capable of driving the vehicle autonomously in urban paths as well as on fast express roads. We describe our whole system, especially novel modules of probabilistic perception fusion, incremental mapping, the 1st and the 2nd planning and the overall safety concern. TiEV finished 2016 and 2017 Intelligent Vehicle Future Challenge of China held at Changshu. We show our experiences on the development of autonomous vehicles and future trends.
  • The gravitational wave (GW) as a standard siren directly determines the luminosity distance from the gravitational waveform without reference to the specific cosmological model, of which the redshift can be obtained separately by means of the electromagnetic counterpart like GW events from binary neutron stars and massive black hole binaries (MBHBs). To see to what extent the standard siren can reproduce the presumed dipole anisotropy written in the simulated data of standard siren events from typical configurations of GW detectors, we find that (1) for the Laser Interferometer Space Antenna with different MBHB models during five-year observations, the cosmic isotropy can be ruled out at $3\sigma$ confidence level (C.L.) and the dipole direction can be constrained roughly around $20\%$ at $2\sigma$ C.L., as long as the dipole amplitude is larger than $0.03$, $0.06$ and $0.025$ for MBHB models Q3d, pop III and Q3nod with increasing constraining ability, respectively; (2) for Einstein Telescope with no less than $200$ standard siren events, the cosmic isotropy can be ruled out at $3\sigma$ C.L. if the dipole amplitude is larger than $0.06$, and the dipole direction can be constrained within $20\%$ at $3\sigma$ C.L. if the dipole amplitude is near $0.1$; (3) for the Deci-Hertz Interferometer Gravitational wave Observatory with no less than $100$ standard siren events, the cosmic isotropy can be ruled out at $3\sigma$ C.L. for dipole amplitude larger than $0.03$ , and the dipole direction can even be constrained within $10\%$ at $3\sigma$ C.L. if the dipole amplitude is larger than $0.07$. Our work manifests the promising perspective of the constraint ability on the cosmic anisotropy from the standard siren approach.
  • In this paper, we propose a fully distributed continuous-time algorithm to solve the distributed optimization problem for second-order multi-agent systems. The optimization objective function is a sum of private cost functions associated to the individual agents and the interaction between agents is described by a weighted undirected graph. We show the exponential convergence of the proposed algorithm if the underlying graph is connected and the private cost functions are strongly convex and have locally Lipschitz gradients. Moreover, to reduce the overall need of communication, we then propose a dynamic event-triggered communication scheme that is free of Zeno behavior. It is shown that the exponential convergence is achieved if the private cost functions are globally Lipschitz. Numerical simulations are provided to illustrate the effectiveness of the theoretical results.
  • Anti-phase and in-phase synchronization modes of a dual-flame system were numerically investigated and theoretically analyzed. The simulated flickering frequencies of a single heptane pool flame with the varying gravity and pool size agree with the previous experiments and the scaling law. Understanding that the deformation, stretch, and pinch-off of a buoyancy-driven flickering flame result from the evolution of toroidal vortices forming from shear layers' roll-up, we analyzed the anti-phase and in-phase synchronization modes in dual pool flames from the perspective of vortex dynamics. The crucial role of the inner-side shear layers between two flames, which resembles the von Karman vortex street in the wake of a bluff body, was identified from the simulation results verified qualitatively against the available experiments in literature. To account for the viscosity effect on vorticity diffusion of the in-phase flickering flames, we defined a flame Reynolds number based on the flow characteristics of the two inner-side shear layers. A unified regime nomogram in the parameter space of the normalized frequency and the Reynolds number was obtained and verified by data from both the current simulation and the previous experiment.
  • Sentiment analysis is a key component in various text mining applications. Numerous sentiment classification techniques, including conventional and deep learning-based methods, have been proposed in the literature. In most existing methods, a high-quality training set is assumed to be given. Nevertheless, constructing a high-quality training set that consists of highly accurate labels is challenging in real applications. This difficulty stems from the fact that text samples usually contain complex sentiment representations, and their annotation is subjective. We address this challenge in this study by leveraging a new labeling strategy and utilizing a two-level long short-term memory network to construct a sentiment classifier. Lexical cues are useful for sentiment analysis, and they have been utilized in conventional studies. For example, polar and privative words play important roles in sentiment analysis. A new encoding strategy, that is, $\rho$-hot encoding, is proposed to alleviate the drawbacks of one-hot encoding and thus effectively incorporate useful lexical cues. We compile three Chinese data sets on the basis of our label strategy and proposed methodology. Experiments on the three data sets demonstrate that the proposed method outperforms state-of-the-art algorithms.
  • For a multiplier Hopf algebra pairing $\langle A, B\rangle$, we construct a class of group-cograded multiplier Hopf algebras $D(A, B)$, generalizing the classical construction of finite dimensional Hopf algebras introduced by Panaite and Staic Mihai. Furthermore, if the multiplier Hopf algebra pairing admits a canonical multiplier in $M(B\otimes A)$ we show the existence of quasitriangular structure on $D(A, B)$. As an application, some special cases and examples are provided.
  • Thermal conductivity of binary aqueous solutions of 1,2-ethanediol and 1,2-propanediol was measured using the transient hot wire method at temperature from 253.15 K to 373.15 K at atmospheric. Measurements were made for six compositions over the entire concentration range from 0 to 1 mole fraction of glycol, namely, 0.0, 0.2, 0.4, 0.6, 0.8, and 1.0 mole fraction of glycol. The uncertainty of the thermal conductivity, temperature, and concentration measurements are estimated to be better than 2%, 0.01K, 0.1%, respectively. The second-order Scheff\'e polynomial is used to correlate the temperature and composition dependence of the experimental thermal conductivity, which is found to be in good agreement with the experiment data from the present work and other reports.
  • Semantic image segmentation is one of the most challenged tasks in computer vision. In this paper, we propose a highly fused convolutional network, which consists of three parts: feature downsampling, combined feature upsampling and multiple predictions. We adopt a strategy of multiple steps of upsampling and combined feature maps in pooling layers with its corresponding unpooling layers. Then we bring out multiple pre-outputs, each pre-output is generated from an unpooling layer by one-step upsampling. Finally, we concatenate these pre-outputs to get the final output. As a result, our proposed network makes highly use of the feature information by fusing and reusing feature maps. In addition, when training our model, we add multiple soft cost functions on pre-outputs and final outputs. In this way, we can reduce the loss reduction when the loss is back propagated. We evaluate our model on three major segmentation datasets: CamVid, PASCAL VOC and ADE20K. We achieve a state-of-the-art performance on CamVid dataset, as well as considerable improvements on PASCAL VOC dataset and ADE20K dataset
  • In this paper, we simulate the Super-Eddington accreting massive black holes (SEAMBHs) as the candles to probe the cosmology for the first time. SEAMBHs have been demonstrated that they can provide a new tool for estimating cosmological distance. Thus we create a series of mock data sets of SEAMBHs specially in hight redshift region to check the abilities of them on probing the cosmology. To fulfill the potential of the SEAMBHs on the cosmology, we apply the simulated data to three projects. The first is exploring their abilities to constrain the cosmological parameters, in which we combine different data sets of current observations such as the cosmic microwave background (CMB) from {\it Planck} and type Ia supernovae (SNe Ia) from Joint Light-curve Analysis (JLA). We find that the high redshift SEAMBHs can help to break the degeneracies of the background cosmological parameters constrained by {\it Planck} and JLA, thus give a much tighter constraints of the cosmological parameters. The second is using the high redshift SEAMBHs as the complements of the low redshift JLA to constrain the early expansion rate. Our results show that these high redshift SEAMBHs are very powerful on constraining the early Hubble rate, thus can give us more information of the expansion history of our Universe, which is also crucial for the test of $\Lambda$CDM model in high redshift region. Finally, we check the SEAMBHs candles' abilities to reconstruct the equation of state of dark energy at high redshift. In summary, our results show that the SEAMBHs, as the rare candles in high redshift region, can provide us a new and independent observation to probe the evolution of the universe in the future.
  • By combining the algebraic Bethe ansatz and the off-diagonal Bethe ansatz, we investigate the trigonometric SU(3) model with generic open boundaries. The eigenvalues of the transfer matrix are given in terms of an inhomogeneous T-Q relation, and the corresponding eigenstates are expressed in terms of nested Bethe-type eigenstates which have well-defined homogeneous limit. This exact solution provides basis for further analyzing the thermodynamic properties and correlation functions of the anisotropic models associated with higher rank algebras.
  • TiO2 nanorods have been prepared on ITO substrates by dc reactive magnetron sputtering technique. The hydroxyl groups have been introduced on the nanorods surface. The structure and the optical properties of these nanorods have been studied. The dye-sensitized solar cells (DSSCs) have been assembled using these TiO2 nanorods as photoelectrode. And the effect of the hydroxyl groups on the properties of the photoelectric conversion of the DSSCs has been studied.
  • We report a new scenario of time-of-flight (TOF) technique in which fast neutrons and delayed gamma-ray signals were both recorded in a millisecond time window in harsh environments induced by high-intensity lasers. The delayed gamma signals, arriving far later than the original fast neutron and often being ignored previously, were identified to be the results of radiative captures of thermalized neutrons. The linear correlation between gamma photon number and the fast neutron yield shows that these delayed gamma events can be employed for neutron diagnosis. This method can reduce the detecting efficiency dropping problem caused by prompt high-flux gamma radiation, and provides a new way for neutron diagnosing in high-intensity laser-target interaction experiments.
  • Genome-wide association studies (GWA studies or GWAS) investigate the relationships between genetic variants such as single-nucleotide polymorphisms (SNPs) and individual traits. Recently, incorporating biological priors together with machine learning methods in GWA studies has attracted increasing attention. However, in real-world, nucleotide-level bio-priors have not been well-studied to date. Alternatively, studies at gene-level, for example, protein--protein interactions and pathways, are more rigorous and legitimate, and it is potentially beneficial to utilize such gene-level priors in GWAS. In this paper, we proposed a novel two-level structured sparse model, called Sparse Group Lasso with Group-level Graph structure (SGLGG), for GWAS. It can be considered as a sparse group Lasso along with a group-level graph Lasso. Essentially, SGLGG penalizes the nucleotide-level sparsity as well as takes advantages of gene-level priors (both gene groups and networks), to identifying phenotype-associated risk SNPs. We employ the alternating direction method of multipliers algorithm to optimize the proposed model. Our experiments on the Alzheimer's Disease Neuroimaging Initiative whole genome sequence data and neuroimage data demonstrate the effectiveness of SGLGG. As a regression model, it is competitive to the state-of-the-arts sparse models; as a variable selection method, SGLGG is promising for identifying Alzheimer's disease-related risk SNPs.
  • The gravitational waves from compact binary systems are viewed as a standard siren to probe the evolution of the universe. This paper summarizes the potential and ability to use the gravitational waves to constrain the cosmological parameters and the dark sector interaction in the Gaussian process methodology. After briefly introducing the method to reconstruct the dark sector interaction by the Gaussian process, the concept of standard sirens and the analysis of reconstructing the dark sector interaction with LISA are outlined. Furthermore, we estimate the constraint ability of the gravitational waves on cosmological parameters with ET. The numerical methods we use are Gaussian process and the Markov-Chain Monte-Carlo. Finally, we also forecast the improvements of the abilities to constrain the cosmological parameters with ET and LISA combined with the \it Planck.
  • We consider a D2D-enabled cellular network where user equipments (UEs) owned by rational users are incentivized to form D2D pairs using tokens. They exchange tokens electronically to "buy" and "sell" D2D services. Meanwhile the devices have the ability to choose the transmission mode, i.e. receiving data via cellular links or D2D links. Thus taking the different benefits brought by diverse traffic types as a prior, the UEs can utilize their tokens more efficiently via transmission mode selection. In this paper, the optimal transmission mode selection strategy as well as token collection policy are investigated to maximize the long-term utility in the dynamic network environment. The optimal policy is proved to be a threshold strategy, and the thresholds have a monotonicity property. Numerical simulations verify our observations and the gain from transmission mode selection is observed.
  • The direct detection of gravitational wave by Laser Interferometer Gravitational-Wave Observatory indicates the coming of the era of gravitational-wave astronomy and gravitational-wave cosmology. It is expected that more and more gravitational-wave events will be detected by currently existing and planned gravitational-wave detectors. The gravitational waves open a new window to explore the Universe and various mysteries will be disclosed through the gravitational-wave detection, combined with other cosmological probes. The gravitational-wave physics is not only related to gravitation theory, but also is closely tied to fundamental physics, cosmology and astrophysics. In this review article, three kinds of sources of gravitational waves and relevant physics will be discussed, namely gravitational waves produced during the inflation and preheating phases of the Universe, the gravitational waves produced during the first-order phase transition as the Universe cools down and the gravitational waves from the three phases: inspiral, merger and ringdown of a compact binary system, respectively. We will also discuss the gravitational waves as a standard siren to explore the evolution of the Universe.
  • We consider the global consensus problem for multi-agent systems with input saturation over digraphs. Under a mild connectivity condition that the underlying digraph has a directed spanning tree, we use Lyapunov methods to show that the widely used distributed consensus protocol, which solves the consensus problem for the case without input saturation constraints, also solves the global consensus problem for the case with input saturation constraints. In order to reduce the overall need of communication and system updates, we then propose a distributed event-triggered control law. Global consensus is still realized and Zeno behavior is excluded. Numerical simulations are provided to illustrate the effectiveness of the theoretical results.
  • We perform a forecast analysis of the ability of the LISA space-based interferometer to reconstruct the dark sector interaction using gravitational wave (GW) standard sirens at high redshift. We employ Gaussian process methods to reconstruct the distance-redshift relation in a model independent way. We adopt simulated catalogues of standard sirens given by merging massive black hole binaries (MBHBs) visible by LISA, with an electromagnetic (EM) counterpart detectable by future telescopes. The catalogues are constructed considering three different astrophysical scenarios for the evolution of MBHB mergers based on the semi-analytic model of E. Barausse (2012). We first use these standard siren datasets to assess the potential of LISA in reconstructing a possible interaction between vacuum dark energy and dark matter. Then we combine the LISA cosmological data with supernovae data simulated for the Dark Energy Survey (DES). We consider two scenarios distinguished by the time duration of the LISA mission: 5 and 10 years. Using only LISA standard siren data, the dark sector interaction can be well reconstructed from redshift $z\sim1$ to $z\sim3$ (5 yr) and $z\sim1$ up to $z\sim5$ (10 yr), though the reconstruction is inefficient at lower redshift. When combined with the DES datasets, the interaction is well reconstructed in the whole redshift region from $z\sim0$ to $z\sim3$ (5 yr) and $z\sim0$ to $z\sim5$ (10 yr). MBHB standard sirens can thus be used to constrain the dark sector interaction at redshift ranges not reachable by usual supernovae datasets which probe only the $z\lesssim 1.5$ range. GW standard sirens will not only constitute a complementary and alternative way, with respect to familiar EM observations, to probe the cosmic expansion, but will also provide new tests to constrain possible deviations from the standard $\Lambda$CDM dynamics, especially at high redshift.
  • We investigate the constraint ability of the gravitational wave (GW) as the standard siren on the cosmological parameters by using the third-generation gravitational wave detector: the Einstein Telescope. We simulate the luminosity distances and redshift measurements from 100 to 1000 GW events. We use two different algorithms to constrain the cosmological parameters. For the Hubble constant $H_0$ and dark matter density parameter $\Omega_m$, we adopt the Markov chain Monte Carlo approach. We find that with about 500-600 GW events we can constrain the Hubble constant with an accuracy comparable to \textit{Planck} temperature data and \textit{Planck} lensing combined results, while for the dark matter density, GWs alone seem not able to provide the constraints as good as for the Hubble constant; the sensitivity of 1000 GW events is a little lower than that of \textit{Planck} data. It should require more than 1000 events to match the \textit{Planck} sensitivity. Yet, for analyzing the more complex dynamical property of dark energy, i.e., the equation of state $w$, we adopt a new powerful nonparametric method: the Gaussian process. We can reconstruct $w$ directly from the observational luminosity distance at every redshift. In the low redshift region, we find that about 700 GW events can give the constraints of $w(z)$ comparable to the constraints of a constant $w$ by \textit{Planck} data with type Ia supernovae. Those results show that GWs as the standard sirens to probe the cosmological parameters can provide an independent and complementary alternative to current experiments.
  • Demand response represents a significant but largely untapped resource that can greatly enhance the flexibility and reliability of power systems. This paper proposes a hierarchical control framework to facilitate the integrated coordination between distributed energy resources and demand response. The proposed framework consists of coordination and device layers. In the coordination layer, various resource aggregations are optimally coordinated in a distributed manner to achieve the system-level objectives. In the device layer, individual resources are controlled in real time to follow the optimal power dispatch signals received from the coordination layer. For practical applications, a method is presented to determine the utility functions of controllable loads by accounting for the real-time load dynamics and the preferences of individual customers. The effectiveness of the proposed framework is validated by detailed simulation studies.
  • We investigate the thermodynamic limit of the su(n)-invariant spin chain models with unparallel boundary fields. It is found that the contribution of the inhomogeneous term in the associated T-Q relation to the ground state energy does vanish in the thermodynamic limit. This fact allows us to calculate the boundary energy of the system. Taking the su(2) (or the XXX) spin chain and the su(3) spin chain as concrete examples, we have studied the corresponding boundary energies of the models. The method used in this paper can be generalized to study the thermodynamic properties and boundary energy of other high rank models with non-diagonal boundary fields.
  • Bell's theorem shows a profound contradiction between local realism and quantum mechanics on the level of statistical predictions. It does not involve directly Einstein-Podolsky-Rosen (EPR) correlations. The paradox of Greenberger-Horne-Zeilinger (GHZ) disproves directly the concept of EPR elements of reality, based on the EPR correlations, in an all-versus-nothing way. A three-qubit experimental demonstration of the GHZ paradox was achieved nearly twenty years ago, and followed by demonstrations for more qubits. Still, the GHZ contradictions underlying the tests can be reduced to three-qubit one. We show an irreducible four-qubit GHZ paradox, and report its experimental demonstration. The reducibility loophole is closed. The bound of a three-setting per party Bell-GHZ inequality is violated by $7\sigma$. The fidelity of the GHZ state was around $81\%$, and an entanglement witness reveals a violation of the separability threshold by $19\sigma$.
  • The quantum $\tau_2$-model with generic site-dependent inhomogeneity and arbitrary boundary fields is studied via the off-diagonal Bethe Ansatz method. The eigenvalues of the corresponding transfer matrix are given in terms of an inhomogeneous T-Q relation, which is based on the operator product identities among the fused transfer matrices and the asymptotic behavior of the transfer matrices. Moreover, the associated Bethe Ansatz equations are also obtained.