arXiv subjects
Publications and source records attributed to Jun Zhang.
The shape invariant of a symplectic manifold encodes the possible area classes of embedded Lagrangian tori. Potentially this is a powerful invariant, but for most manifolds the shape is unknown. We compute the shape for 4 dimensional symplectic ellipsoids where the ratio of the factors is an integer. The full shape invariant gives stronger embedding obstructions than results by considering only monotone tori.
In this paper, we generalize the result from L. Polterovich and E. Shelukhin's paper stating that Hofer distance from time-dependent Hamiltonian diffeomorphism to the set of p-th power Hamiltonian diffeomorphism can be arbitrarily large to hold in the product structure $Σ_g \times M$ for any closed symplectic manifold $M$ when $p$ is sufficiently large and $g \geq 4$. This implies that, on this product, Hofer distance can be arbitrarily large between time-dependent Hamiltonian diffeomorphism and autonomous Hamiltonian diffeomorphism.The basic tool we use is barcode and singular value decomposition that are developed in previous joint work with M. Usher, from which we borrow many proofs and modify them so that it can be adapted to the situation that filtered chain complex equipped with a group action.
We report the synthesis of a new quasi one-dimensional (1D) iron selenide. Ba9Fe3Se15 was synthesized at high temperature and high pressure of 5.5 GPa and systematically studied via structural, magnetic and transport measurements at ambient and at high-pressures. Ba9Fe3Se15 crystallizes in a monoclinic structure and consists of face-sharing FeSe6 octahedral chains along the c axis. At ambient pressure it exhibits an insulating behavior with a band gap ~460 meV and undergoes a ferrimagnet-like phase transition at 14 K. Under high pressure, a complete metallization occurs at ~29 GPa, which is accompanied by a spin state crossover from high spin (HS) state to low spin (LS) state. The LS appears for pressures P >36 GPa.
The theory of persistence modules is an emerging field of algebraic topology which originated in topological data analysis. In these notes we provide a concise introduction into this field and give an account on some of its interactions with geometry and analysis. In particular, we present applications of persistence to symplectic topology, including the geometry of symplectomorphism groups and embedding problems. Furthermore, we discuss topological function theory which provides a new insight on oscillation of functions. The material should be accessible to readers with a basic background in algebraic and differential topology.
The acoustic modes of diamond not only are of profound significance for studying its thermal conductivity, mechanical properties and optical properties, but also play a determined role in the performance of high-frequency and high-power acoustic wave devices. Here we report the bulk acoustic waves (BAWs) and surface acoustic waves (SAWs) of single crystal diamond by using the angle-resolved Brillouin light scattering (BLS) spectroscopy. We identify two high-velocity surface skimming bulk waves, with sound velocities of 1.277x10^6 and 1.727x10^6 cm/s, respectively. Furthermore, we conduct the relationship among the refractive index, incident angle and the velocities of BAWs propagating along an arbitrary direction. Our results may provide a valuable reference for fundamental researches and devices engineering in the community of diamond-based acoustic study.
Text-based person search aims at retrieving target person in an image gallery using a descriptive sentence of that person. It is very challenging since modal gap makes effectively extracting discriminative features more difficult. Moreover, the inter-class variance of both pedestrian images and descriptions is small. So comprehensive information is needed to align visual and textual clues across all scales. Most existing methods merely consider the local alignment between images and texts within a single scale (e.g. only global scale or only partial scale) then simply construct alignment at each scale separately. To address this problem, we propose a method that is able to adaptively align image and textual features across all scales, called NAFS (i.e.Non-local Alignment over Full-Scale representations). Firstly, a novel staircase network structure is proposed to extract full-scale image features with better locality. Secondly, a BERT with locality-constrained attention is proposed to obtain representations of descriptions at different scales. Then, instead of separately aligning features at each scale, a novel contextual non-local attention mechanism is applied to simultaneously discover latent alignments across all scales. The experimental results show that our method outperforms the state-of-the-art methods by 5.53% in terms of top-1 and 5.35% in terms of top-5 on text-based person search dataset. The code is available at https://github.com/TencentYoutuResearch/PersonReID-NAFS
Strategy-proof mechanisms are widely used in market design. In an abstract allocation framework where outside options are available to agents, we obtain two results for strategy-proof mechanisms. They provide a unified foundation for several existing results in distinct models and imply new results in some models. The first result proves that, for individually rational and strategy-proof mechanisms, pinning down every agent's probability of choosing his outside option is equivalent to pinning down a mechanism. The second result provides a sufficient condition for two strategy-proof mechanisms to be equivalent when the number of possible allocations is finite.
Sub-percent level accuracy in shear measurement is required by the Stage-IV weak lensing surveys. One important challenge is about suppressing the shear bias on source images of low signal-to-noise ratios (SNR$\lesssim10$). Previously, it has been demonstrated that the shear estimators defined in the Fourier_Quad (FQ) method can achieve sub-percent accuracy at the very faint end (SNR$\lesssim5$) through ensemble averaging. Later, it is found that we can approach the minimum statistical error (the Cramer-Rao Bound) by symmetrizing the full PDF of the FQ shear estimators (the PDF_SYM approach), instead of taking ensemble averages. Recently, with a large amount of mock galaxy images, we are able to identify some small amount of shear biases in the PDF_SYM approach at the faint end. The multiplicative bias goes up to $1-2\times10^{-2}$ at SNR $\lesssim 10$, and the anisotropy of the point spread function (PSF) causes an additive bias that can reach a few times $10^{-4}$. We find that these biases originate from the noise-source coupling in the galaxy power spectrum. It turns out that this problem can be largely fixed by adding additional terms to the FQ shear estimators. The resulting multiplicative and additive biases can be significantly suppressed to the level of $10^{-3}$ and $10^{-5}$ respectively. These corrections substantially extend the available SNR range for accurate shear measurement with the PDF_SYM approach.
Massive MIMO has been regarded as a key enabling technique for 5G and beyond networks. Nevertheless, its performance is limited by the large overhead needed to obtain the high-dimensional channel information. To reduce the huge training overhead associated with conventional pilot-aided designs, we propose a novel blind data detection method by leveraging the channel sparsity and data concentration properties. Specifically, we propose a novel $\ell_3$-norm-based formulation to recover the data without channel estimation. We prove that the global optimal solution to the proposed formulation can be made arbitrarily close to the transmitted data up to a phase-permutation ambiguity. We then propose an efficient parameter-free algorithm to solve the $\ell_3$-norm problem and resolve the phase permutation ambiguity. We also derive the convergence rate in terms of key system parameters such as the number of transmitters and receivers, the channel noise power, and the channel sparsity level. Numerical experiments will show that the proposed scheme has superior performance with low computational complexity.
In image processing, source detections are inevitably affected by the presence of the geometric boundaries in the images, including the physical boundaries of the CCD, and the boundaries of masked regions due to column defects, bright diffraction spikes, etc.. These boundary conditions make the source detection process not statistically isotropic. It can lead to additive shear bias near the boundaries. We build a phenomenological model to understand the bias, and propose a simple method to effectively eliminate the influence of geometric boundaries on shear measurement. We demonstrate the accuracy and efficiency of this method using both simulations and the z-band imaging data from the third data release of the DECam Legacy Survey.
We present the formation and eruption of a sigmoidal filament driven by rotating network magnetic fields (RNFs) near the center of the solar disk, which was observed by the one-meter aperture New Vacuum Solar Telescope (NVST) at Fuxian Solar Observatory (FSO) on 2018 July 12. Counterclockwise RNFs twist two small-scale filaments at their northeastern foot-point region, giving a rotation of nearly 200 degree within about 140 minutes. The motion of the RNF has a tendency to accelerate at first and then decelerate obviously, as the average rotation speed increased from 10 to 150 ,and then slowed down to 50 . Coalescence then occurs between filaments F1 and F2. Meanwhile the fine structures in the southwestern region of the filament was involved in another interaction of coalescence. The subsequent EUV brightening due to plasma heating is observed in the two interaction regions. These interacting structures, including F1, F2 and the fine structures in the southwestern region, eventually evolve into a larger-scale sigmoidal filament twisted in the same direction as the RNFs gave. The twist of the sigmoidal filament has exceeded 4π and the filament erupted finally. The motion of the sigmoidal filament keeps uniform until a nearby jet collides, causing the filament to erupt faster. These results provide evidence that RNF plays an important role in the formation and eruption of the sigmoidal filament. The phenomena also suggests that the kink instability is the trigger mechanism for the filament eruption.
User modeling is critical for developing personalized services in industry. A common way for user modeling is to learn user representations that can be distinguished by their interests or preferences. In this work, we focus on developing universal user representation model. The obtained universal representations are expected to contain rich information, and be applicable to various downstream applications without further modifications (e.g., user preference prediction and user profiling). Accordingly, we can be free from the heavy work of training task-specific models for every downstream task as in previous works. In specific, we propose Self-supervised User Modeling Network (SUMN) to encode behavior data into the universal representation. It includes two key components. The first one is a new learning objective, which guides the model to fully identify and preserve valuable user information under a self-supervised learning framework. The other one is a multi-hop aggregation layer, which benefits the model capacity in aggregating diverse behaviors. Extensive experiments on benchmark datasets show that our approach can outperform state-of-the-art unsupervised representation methods, and even compete with supervised ones.
Silicon single-photon detectors (SPDs) are key devices for detecting single photons in the visible wavelength range. Photon detection efficiency (PDE) is one of the most important parameters of silicon SPDs, and increasing PDE is highly required for many applications. Here, we present a practical approach to increase PDE of silicon SPD with a monolithic integrated circuit of active quenching and active reset (AQAR). The AQAR integrated circuit is specifically designed for thick silicon single-photon avalanche diode (SPAD) with high breakdown voltage (250-450 V), and then fabricated via the process of high-voltage 0.35-$μ$m bipolarCMOS-DMOS. The AQAR integrated circuit implements the maximum transition voltage of ~ 68 V with 30 ns quenching time and 10 ns reset time, which can easily boost PDE to the upper limit by regulating the excess bias up to a high enough level. By using the AQAR integrated circuit, we design and characterize two SPDs with the SPADs disassembled from commercial products of single-photon counting modules (SPCMs). Compared with the original SPCMs, the PDE values are increased from 68.3% to 73.7% and 69.5% to 75.1% at 785 nm, respectively, with moderate increases of dark count rate and afterpulse probability. Our approach can effectively improve the performance of the practical applications requiring silicon SPDs.
Combining reinforcement learning (RL) and molecular dynamics (MD) simulations, we propose a machine-learning approach (RL$^‡$) to automatically unravel chemical reaction mechanisms. In RL$^‡$, locating the transition state of a chemical reaction is formulated as a game, where a virtual player is trained to shoot simulation trajectories connecting the reactant and product. The player utilizes two functions, one for value estimation and the other for policy making, to iteratively improve the chance of winning this game. We can directly interpret the reaction mechanism according to the value function. Meanwhile, the policy function enables efficient sampling of the transition paths, which can be further used to analyze the reaction dynamics and kinetics. Through multiple experiments, we show that RL‡ can be trained tabula rasa hence allows us to reveal chemical reaction mechanisms with minimal subjective biases.
We propose a pseudo-market solution to resource allocation problems subject to constraints. Our treatment of constraints is general: including bihierarchical constraints due to considerations of diversity in school choice, or scheduling in course allocation; and other forms of constraints needed to model, for example, the market for roommates, and combinatorial assignment problems. Constraints give rise to pecuniary externalities, which are internalized via prices. Agents pay to the extent that their purchases affect the value of relevant constraints at equilibrium prices. The result is a constrained efficient market equilibrium outcome. The outcome is fair whenever the constraints do not single out individual agents. Our result can be extended to economies with endowments, and address participation constraints.
Group buying, as an emerging form of purchase in social e-commerce websites, such as Pinduoduo, has recently achieved great success. In this new business model, users, initiator, can launch a group and share products to their social networks, and when there are enough friends, participants, join it, the deal is clinched. Group-buying recommendation for social e-commerce, which recommends an item list when users want to launch a group, plays an important role in the group success ratio and sales. However, designing a personalized recommendation model for group buying is an entirely new problem that is seldom explored. In this work, we take the first step to approach the problem of group-buying recommendation for social e-commerce and develop a GBGCN method (short for Group-Buying Graph Convolutional Network). Considering there are multiple types of behaviors (launch and join) and structured social network data, we first propose to construct directed heterogeneous graphs to represent behavioral data and social networks. We then develop a graph convolutional network model with multi-view embedding propagation, which can extract the complicated high-order graph structure to learn the embeddings. Last, since a failed group-buying implies rich preferences of the initiator and participants, we design a double-pairwise loss function to distill such preference signals. We collect a real-world dataset of group-buying and conduct experiments to evaluate the performance. Empirical results demonstrate that our proposed GBGCN can significantly outperform baseline methods by 2.69%-7.36%. The codes and the dataset are released at https://github.com/Sweetnow/group-buying-recommendation.
Distributed quantum sensing can provide quantum-enhanced sensitivity beyond the shot-noise limit (SNL) for sensing spatially distributed parameters. To date, distributed quantum sensing experiments have been mostly accomplished in laboratory environments without a real space separation for the sensors. In addition, the post-selection is normally assumed to demonstrate the sensitivity advantage over the SNL. Here, we demonstrate distributed quantum sensing in field and show the unconditional violation (without post-selection) of SNL up to 0.916 dB for the field distance of 240 m. The achievement is based on a loophole free Bell test setup with entangled photon pairs at the averaged heralding efficiency of 73.88%. Moreover, to test quantum sensing in real life, we demonstrate the experiment for long distances (with 10-km fiber) together with the sensing of a completely random and unknown parameter. The results represent an important step towards a practical quantum sensing network for widespread applications.
In this work we propose an inference technique, asynchronous revision, to unify streaming and non-streaming speech recognition models. Specifically, we achieve dynamic latency with only one model by using arbitrary right context during inference. The model is composed of a stack of convolutional layers for audio encoding. In inference stage, the history states of encoder and decoder can be asynchronously revised to trade off between the latency and the accuracy of the model. To alleviate training and inference mismatch, we propose a training technique, segment cropping, which randomly splits input utterances into several segments with forward connections. This allows us to have dynamic latency speech recognition results with large improvements in accuracy. Experiments show that our dynamic latency model with asynchronous revision gives 8\%-14\% relative improvements over the streaming models.