Searcharxiv⌕ Search

arXiv subjects

Jian Wang

Publications and source records attributed to Jian Wang.

At least 901 records · Page 50Linked to original sources
arXiv2019

Improving Knowledge-aware Dialogue Generation via Knowledge Base Question Answering

Neural network models usually suffer from the challenge of incorporating commonsense knowledge into the open-domain dialogue systems. In this paper, we propose a novel knowledge-aware dialogue generation model (called TransDG), which transfers question representation and knowledge matching abilities from knowledge base question answering (KBQA) task to facilitate the utterance understanding and factual knowledge selection for dialogue generation. In addition, we propose a response guiding attention and a multi-step decoding strategy to steer our model to focus on relevant features for response generation. Experiments on two benchmark datasets demonstrate that our model has robust superiority over compared methods in generating informative and fluent dialogues. Our code is available at https://github.com/siat-nlp/TransDG.

cs.CL↗
arXiv2019

Structured light interferometry

Optical interferometry has made tremendous development in fundamentals and applications of light. Traditional optical interferometry using simplex plane phase light just allows one dimensional measurement, but incapable of retrieving multidimensional movement. Here, we present new structured light interferometry to achieve simultaneous determination of rotational and translational dimensionalities (magnitudes and directions) of moving objects by fully exploiting multiple degree of freedoms of structured light. Such multidimensional detectability benefits from multiple amplitude and sign distinguishable Doppler shifts of structured light by analyzing relative amplitude and phase information of Doppler signals. We experimentally demonstrate two interferometric schemes to individually retrieve both translational and rotational velocity vectors of two kinds of moving objects (small particles and bulk targets). These interferometric schemes based on structured light show remarkable function enhancement for simultaneous measurement of multidimensional motion vectors, compared with traditional interferometry, and thus may expand optical interferometry to multidimensional metrology in remote sensing, intelligent manufacturing and engineering.

physics.optics↗
arXiv2019

A vectorial Doppler effect with spatially variant polarized light fields

The Doppler effect of light was implemented by interference with a reference wave to infer linear velocities in early manifestations, and more recently lateral and angular velocities with scalar phase structured light. A consequence of the scalar wave approach is that it is impossible to deduce the motion direction and position of moving targets directly. Here we overcome this limitation with vectorially structured light, exhibiting spatially variant polarization, allowing both the position and velocity of a particle to be robustly determined by a new vectorial Doppler effect. We use this to successfully implement real time monitoring of instantaneous angular position and velocity of a complicated moving particle, and interpret its physical mechanism as evidence of time reversal symmetry breaking in classically entangled light fields. Our demonstration can easily be extended with arbitrary vector structured fields for complete motion tracking of nontrivial trajectories, offering new vectorial Doppler metrology for universal motion vectors.

physics.optics↗
arXiv2019

Sharing a common origin between the rotational and linear Doppler effects

The well-known linear Doppler effect arises from the linear motion between source and observer, while the less well-known rotational Doppler effect arises from the rotational motion. Here, we present both theories and experiments illustrating the relationship between the rotational and linear Doppler effects. A spiral phaseplate is used to generate a light beam carrying orbital angular momentum and the frequency shift is measured arising from its rotational and/or linear motion. By considering either the motion-induced time-evolving phase or the momentum and energy conservation in light-matter interactions, we derive the rotational Doppler shift, linear Doppler shift, and overall Doppler shift associated with rotational and linear motions. We demonstrate the relationship between rotational and linear Doppler shifts, either of which can be derived from the other effect, thereby illustrating their shared origin. Moreover, the close relationship between rotational and linear Doppler effects is also deduced for a more general moving rough surface.

physics.optics↗
arXiv2019

Intrinsic transverse spin angular momentum of fiber eigenmodes

We study the transverse spin angular momentum of fiber eigenmodes. As an intrinsic property of fiber eigenmodes, it becomes considerable in optical nanofibers and, especially, increases sharply in the evanescent field. We derive analytical expressions of this angular momentum, and present its density distribution inside andoutside nanofibers. Significantly, we find the optimal ratio of the fiber core radius to the wavelength to obtain the maximum surface transverse spin. For instance, the optimal fiber radius is one fourth of the wavelength for the fundamental mode. Furthermore, we investigate the spin flow of each guided mode on the nanofiber surface and the transverse mechanical effects originated from the transverse spin on small particles.

physics.optics↗
arXiv2019

Mode conversion and orbital angular momentum transfer among multiple modes by helical gratings

The coupling feature of helical gratings (HGs) has been studied on flexible conversion between two orbital angular momentum (OAM) modes by both the transverse and longitudinal modulation of HGs. Apart from one to one OAM exchange, HGs can achieve OAM transfer among multiple modes, provided both the helix and phase matching conditions are well satisfied for successive coupling. It is contributed from the fold number of modulation fringes of HGs and the difference multiplication of propagation constants between two of the successive modes. Based on the coupled mode theory, we investigate the mode conversion and OAM transfer among three modes using a 3 fold HGs in a ring core fiber, and numerically simulate it by transmission matrix. Our work about the multiple modes transferring can be regarded as an extension of mode coupling for optical gratings, and might flourish the mode conversion, as well as OAM manipulation in optics fields.

physics.optics↗
arXiv2019

Full-Vectorial Mode Coupling in Optical Fibers

Optical full vectorial modes have closely attracted much attention because of their unique optical features and thus may develop the fundamental researches and new functionalities of light. Beyond the conventional linear polarization modal coupling, we thoroughly investigate the inherent polarizationdependence and polarization-isolation features for full vectorial mode coupling in optical fibers, as well as the coupling crosstalk in view of the modal degeneracy. We present the general full vectorial coupling mechanism and numerically simulate the full vectorial mode coupling for both fiber directional and grating assisted coupling cases. Furthermore, we give the detailed analyses of coupling crosstalk based on the unified threemode coupling model. Apart from lifting the modal degeneracy, significantly, the crosstalk can also be reduced by weakening the coupling coefficient and meanwhile increasing the coupling length. This extended full vectorial mode coupling mechanismmay provide theoretical guidance on exploiting new all fiberbased polarization-sensitive passive elements to control fullvectorial fields and modal polarization states.

physics.optics↗
arXiv2019

Optical angular momentum derivation and evolution from vectorial field superposition

Optical intrinsic angular momentum can be regarded as derivation from spatial superposition of optical vector fields embodied by spinning or and spiraling the electric field vector. We employ vectorial formulation derivation to comprehensively study all angular momentum contents of optical vector fields in arbitrary superposition states, including the longitudinal and transverse, spin and orbital (SAM and OAM) components. As for the orthogonal superposition fields, there inherently exists spin-orbit shift from longitudinal SAM to OAM, and the whole local spin flow manifests local multiple fold helical trajectories. Especially, both the spin-orbit shift and transverse SAM could become considerable in the non-paraxial condition. Our studies here provide an explicit insight into the derivation and evolution, intrinsic correlations and salient features of various types of angular momentum components.

physics.optics↗
arXiv2019

All-fiber polarization-manageable directional couplers

Flexible management of light polarization is necessary to develop the technologies of optical fiber communications. However, it is still a great challenge to manipulate optical polarization in all-fiberbased devices. Here we present a novel structural design of three-core fiber couplers to create arbitrary polarization management by embedding a vector mode segregated core within the conventional fiber couplers. We take advantages of these coupling structures to achieve polarization split, rotation, and exchange, as three essential polarization managements for fiber-guided modes. We study these couplers based on coupled-mode theory and numerically simulations. The obtained results show that they manifest favorable operation performance on management of the fiber modal polarization states. Compared to previous couplers used for management of power, wavelengths and modal forms, this kind of fiber couplers introduces a new manageable degree of freedom for modal polarization. It is expected that the presented all-fiber polarization-manageable directional couplers might find potential applications in fiber optical communications and sensing systems.

physics.optics↗
arXiv2019

Flexible generation conversion and exchange of fiber guided orbital angular momentum modes using helical gratings

By exploiting helical gratings (HGs), we propose and simulate flexible generation, conversion and exchange of fiber guided orbital angular momentum (OAM) modes. HGs can enable the generation of OAM modes, and the OAM conversion between two arbitrary modes guided in fibers. A specific HG can exchange the OAM states of a couple of OAM modes, i.e, OAM exchange. In addition, a Fabry Perot cavity cascaded with two identical reflective HGs can reflect converted OAM modes with a comb spectrum. The HGs based generation conversion and exchange of OAM modes are dependent on helix period, orientation, and the fold number of helical fringes. The proposed method of generation, conversion, and exchange of fiber guided OAM modes using HGs is flexible and well compatible with OAM fibers, featuring a high conversion efficiency close to 100% and a conversion bandwidth about 10 nm in transmission spectra, while less than 1 nm in reflection spectra.

physics.optics↗
arXiv2019

Ultra-directional high-efficient chiral silicon photonic circuits

Chiral light matter interaction enables new fundamental researches and applications of light. The interaction has traditionally faced challenges in low directionality and efficiency based on spin orbit interaction of light in microscopic waveguides. It is pivotal to exploit photonic integrated circuits to efficiently engineer photonic chiral behavior. Here, we present ultra directional high efficient chiral coupling in silicon photonic circuits based on low order to high order mode conversion and interference. We show that the directionality of chiral coupling, in principle, can approach minus/plus 1 with circular polarization inputs, benefited from the underlying mechanism of complete destructive and constructive interference. The chiral coupling efficiency can exceed 70%, with negligible scattering to nonguided modes, much higher than conventional coupling mechanisms. Moreover, the chiral silicon photonic circuits can function as a perfect 3 dB power splitter for arbitrarily linear polarization inputs, and also open up the possibility of on chip chirality determination to further flourish the development of chiral optics.

physics.app-ph↗
arXiv2019

Exact solution of a anisotropic $J_1-J_2$ spin chain with antiperiodic boundary condition

The exact solution of an integrable anisotropic Heisenberg spin chain with nearest-neighbour, next-nearest-neighbour and scalar chirality couplings is studied, where the boundary condition is the antiperiodic one. The detailed construction of Hamiltonian and the proof of integrability are given. The antiperiodic boundary condition breaks the $U(1)$-symmetry of the system and we use the off-diagonal Bethe Ansatz to solve it. The energy spectrum is characterized by the inhomogeneous $T-Q$ relations and the contribution of the inhomogeneous term is studied. The ground state energy and the twisted boundary energy in different regions are obtained. We also find that the Bethe roots at the ground state form the string structure if the coupling constant $J=-1$ although the Bethe Ansatz equations are the inhomogeneous ones.

math-ph↗
arXiv2019

An Experimentally Verified Approach to non-Entanglement-Breaking Channel Certification

Ensuring the non-entanglement-breaking (non-EB) property of quantum channels is crucial for the effective distribution and storage of quantum states. However, a practical method for direct and accurate certification of the non-EB feature is highly desirable. Here, we propose and verify a realistic source based measurement device independent certification of non-EB channels. Our method is resilient to repercussions on the certification from experimental conditions, such as multiphotons and imperfect state preparation, and can be implemented with information incomplete set. We achieve good agreement between experimental outcomes and theoretical predictions, which is validated by the expected results of the ideal semi-quantum signaling game, and accurately certify the non-EB channels. Furthermore, our approach is highly robust to effects from noise. Therefore, the proposed approach can be expected to play a significant role in the design and evaluation of realistic quantum channels.

quant-ph↗
arXiv2019

Predicting CMB dust foreground using galactic 21 cm data

Understanding large-angular-scale galactic foregrounds is crucial for future CMB experiments aiming to detect $B$-mode polarization from primordial gravitational waves. Traditionally, the dust component has been separated using its different frequency dependence. However, using non-CMB observations has potential to increase fidelity and decrease the reconstruction noise. In this exploratory paper we investigate the capability of galactic 21 cm observations to predict the dust foreground in intensity. We train a neural network to predict the dust foreground as measured by the Planck Satellite from the full velocity data-cube of galactic 21 cm emission as measured by the HI4PI survey. We demonstrate that information in the velocity structure clearly improves the predictive power over both a simple integrated emission model and a simple linear model. The improvement is significant at arc-minute scales but more modest at degree scales. This proof of principle on temperature data indicates that it might also be possible to improve foreground polarization templates from the same input data.

astro-ph.CO↗
arXiv2019

Fast Color Constancy with Patch-wise Bright Pixels

In this paper, a learning-free color constancy algorithm called the Patch-wise Bright Pixels (PBP) is proposed. In this algorithm, an input image is first downsampled and then cut equally into a few patches. After that, according to the modified brightness of each patch, a proper fraction of brightest pixels in the patch is selected. Finally, Gray World (GW)-based methods are applied to the selected bright pixels to estimate the illuminant of the scene. Experiments on NUS $8$-Camera Dataset show that the PBP algorithm outperforms the state-of-the-art learning-free methods as well as a broad range of learning-based ones. In particular, PBP processes a $1080$p image within two milliseconds, which is hundreds of times faster than the existing learning-free ones. Our algorithm offers a potential solution to the full-screen smart phones whose screen-to-body ratio is $100$\%.

cs.CV↗
arXiv2019

The shifting method and generalized Turán number of matchings

Given two graphs $T$ and $F$, the maximum number of copies of $T$ in an $F$-free graph on $n$ vertices is called the generalized Turán number, denoted by $ex(n,T,F)$. When $T=K_2$, it reduces to the classical Turán number $ex(n,F)$. Let $M_{k}$ be a matching with $k$ edges and $K^{*}_{s,t}$ a graph obtained from $K_{s,t}$ by replacing the part of size $s$ by a clique of the same size. In this paper, we show that for any $s\geq 2$ and $n\geq 2k+1$, \[ ex(n,K_s,M_{k+1})=\max\left\{\binom{2k+1}{s}, \binom{k}{s}+(n-k)\binom{k}{s-1}\right\}. \] For any $s\geq 1$, $t\geq 2$ and $n\geq 2k+1$, \[ ex(n,K_{s,t}^*,M_{k+1})=\max\left\{\binom{2k+1}{s+t}\binom{s+t}{t}, \binom{k}{s}\binom{n-s}{t}+(n-k)\binom{k}{s+t-1}\binom{s+t-1}{t}\right\}. \] Moreover, we also study the bipartite case of the problem. Let $ex_{bip}(n,T,F)$ be the maximum possible number of copies of $T$ in an $F$-free bipartite graph with each part of size $n$. We prove that for any $s,t\geq 1$ and $n\geq k$, \[ ex_{bip}(n,K_{s,t},M_{k+1})=\left\{ \begin{aligned} &\binom{k}{s}\binom{n}{t}+\binom{k}{t}\binom{n}{s}, & \quad s\neq t, &\binom{k}{s}\binom{n}{s},&\quad s=t. \end{aligned} \right. \] Our proof is mainly based on the shifting method.

math.CO↗
arXiv2019

Buffer-aware Wireless Scheduling based on Deep Reinforcement Learning

In this paper, the downlink packet scheduling problem for cellular networks is modeled, which jointly optimizes throughput, fairness and packet drop rate. Two genie-aided heuristic search methods are employed to explore the solution space. A deep reinforcement learning (DRL) framework with A2C algorithm is proposed for the optimization problem. Several methods have been utilized in the framework to improve the sampling and training efficiency and to adapt the algorithm to a specific scheduling problem. Numerical results show that DRL outperforms the baseline algorithm and achieves similar performance as genie-aided methods without using the future information.

cs.IT↗