arXiv subjects

Wei Wu

Publications and source records attributed to Wei Wu.

At least 487 records · Page 27Linked to original sources

arXiv2019

Feedback Network for Image Super-Resolution

Recent advances in image super-resolution (SR) explored the power of deep learning to achieve a better reconstruction performance. However, the feedback mechanism, which commonly exists in human visual system, has not been fully exploited in existing deep learning based image SR methods. In this paper, we propose an image super-resolution feedback network (SRFBN) to refine low-level representations with high-level information. Specifically, we use hidden states in an RNN with constraints to achieve such feedback manner. A feedback block is designed to handle the feedback connections and to generate powerful high-level representations. The proposed SRFBN comes with a strong early reconstruction ability and can create the final high-resolution image step by step. In addition, we introduce a curriculum learning strategy to make the network well suitable for more complicated tasks, where the low-resolution images are corrupted by multiple types of degradation. Extensive experimental results demonstrate the superiority of the proposed SRFBN in comparison with the state-of-the-art methods. Code is avaliable at https://github.com/Paper99/SRFBN_CVPR19.

cs.CV↗

arXiv2019

Maximum of the Ginzburg-Landau fields

We study two dimensional massless field in a box with potential $V\left( \nabla ϕ\left( \cdot \right) \right) $ and zero boundary condition, where $V$ is any symmetric and uniformly convex function. Naddaf-Spencer and Miller proved the macroscopic averages of this field converge to a continuum Gaussian free field. In this paper we prove the distribution of local marginal $ϕ\left( x\right) $, for any $x$ in the bulk, has a Gaussian tail. We further characterize the leading order of the maximum and dimension of high points of this field, thus generalize the results of Bolthausen-Deuschel-Giacomin and Daviaud for the discrete Gaussian free field.

math.PR↗

arXiv2019

Neural Response Generation with Meta-Words

We present open domain response generation with meta-words. A meta-word is a structured record that describes various attributes of a response, and thus allows us to explicitly model the one-to-many relationship within open domain dialogues and perform response generation in an explainable and controllable manner. To incorporate meta-words into generation, we enhance the sequence-to-sequence architecture with a goal tracking memory network that formalizes meta-word expression as a goal and manages the generation process to achieve the goal with a state memory panel and a state controller. Experimental results on two large-scale datasets indicate that our model can significantly outperform several state-of-the-art generation models in terms of response relevance, response diversity, accuracy of one-to-many modeling, accuracy of meta-word expression, and human evaluation.

cs.CL↗

arXiv2019

A Document-grounded Matching Network for Response Selection in Retrieval-based Chatbots

We present a document-grounded matching network (DGMN) for response selection that can power a knowledge-aware retrieval-based chatbot system. The challenges of building such a model lie in how to ground conversation contexts with background documents and how to recognize important information in the documents for matching. To overcome the challenges, DGMN fuses information in a document and a context into representations of each other, and dynamically determines if grounding is necessary and importance of different parts of the document and the context through hierarchical interaction with a response at the matching step. Empirical studies on two public data sets indicate that DGMN can significantly improve upon state-of-the-art methods and at the same time enjoys good interpretability.

cs.CL↗

arXiv2019

Learning a Matching Model with Co-teaching for Multi-turn Response Selection in Retrieval-based Dialogue Systems

We study learning of a matching model for response selection in retrieval-based dialogue systems. The problem is equally important with designing the architecture of a model, but is less explored in existing literature. To learn a robust matching model from noisy training data, we propose a general co-teaching framework with three specific teaching strategies that cover both teaching with loss functions and teaching with data curriculum. Under the framework, we simultaneously learn two matching models with independent training sets. In each iteration, one model transfers the knowledge learned from its training set to the other model, and at the same time receives the guide from the other model on how to overcome noise in training. Through being both a teacher and a student, the two models learn from each other and get improved together. Evaluation results on two public data sets indicate that the proposed learning approach can generally and significantly improve the performance of existing matching models.

cs.CL↗

arXiv2019

Improving Matching Models with Hierarchical Contextualized Representations for Multi-turn Response Selection

In this paper, we study context-response matching with pre-trained contextualized representations for multi-turn response selection in retrieval-based chatbots. Existing models, such as Cove and ELMo, are trained with limited context (often a single sentence or paragraph), and may not work well on multi-turn conversations, due to the hierarchical nature, informal language, and domain-specific words. To address the challenges, we propose pre-training hierarchical contextualized representations, including contextual word-level and sentence-level representations, by learning a dialogue generation model from large-scale conversations with a hierarchical encoder-decoder architecture. Then the two levels of representations are blended into the input and output layer of a matching model respectively. Experimental results on two benchmark conversation datasets indicate that the proposed hierarchical contextualized representations can bring significantly and consistently improvement to existing matching models for response selection.

cs.CL↗

arXiv2019

An analytical variational method for the biased quantum Rabi model in the ultra-strong coupling regime

An analytical variational method for the ground state of the biased quantum Rabi model in the ultra-strong coupling regime is presented. This analytical variational method can be obtained by a unitary transformation or alternatively by assuming the form of ground state wave function. The key point of the method is to introduce a variational parameter $λ$, which can be determined by minimizing the energy functional. Using this method, we calculate physical observables with high accuracy in comparison with the numerical exact one. Our method evidently improves the results from the widely used general rotating-wave approximation (GRWA) in both qualitative and quantitative ways.

quant-ph↗

arXiv2019

Variational generalized rotating-wave approximation in the two-qubit quantum Rabi model

We present an analytical method for the two-qubit quantum Rabi model. While still operating in the frame of the generalized rotating-wave approximation (GRWA), our method further embraces the idea of introducing variational parameters. The optimal value of the variational parameter is determined by minimizing the energy function of the ground state. Comparing with numerical exact results, we show that our method evidently improves the accuracy of the conventional GRWA in calculating fundamental physical quantities, such as energy spectra, mean photon number, and dynamics. Interestingly, the accuracy of our method allows us to reproduce the asymptotic behavior of mean photon number in large frequency ratio for the ground state and investigate the quasi-periodical structure of the time evolution, which are incapable of being predicted by the GRWA. The applicable parameter ranges cover the ultrastrong coupling regime, which will be helpful to recent experiments.

quant-ph↗

arXiv2019

Resource Allocation Algorithm for V2X communications based on SCMA

In this paper, we propose a resource allocation algorithm for V2X communications based on Sparse Code Multiple Access(SCMA). By analyzing the interference model in the V2X scenario, we formulate the problem which deals with resource allocation to maximize the system throughput. A graph color-based user cluster algorithm combined with resource allocation algorithm based on both result of clustering and SINR is presented to solve the problem. The simulation results indicate that the throughput performance of system based on SCMA is superior to which based on OFDMA, and the proposed algorithm can improve the system throughput and the number of access users.

eess.SP↗

arXiv2019

Quantum simulation of the Weyl equation with a trapped ion

The Weyl equation describes chiral massless relativistic particles, called Weyl fermions, which have important relations to neutrinos. A direct observation of the dynamics of Weyl fermions in an experiment is difficult to achieve. This study investigates a method of simulating the Weyl equation in 1+2-dimension by a single trapped ion. The predictions about a two-dimensional Zitterbewegung and an especially interesting phenomenon of Weyl fermions can be tested by the future trapped ion experiment, which might enhance our understanding of neutrinos.

quant-ph↗

arXiv2019

Steady-state entanglement and coherence of the coupled qubit system in equilibrium and nonequilibrium environments

We investigate analytically and numerically the steady-state entanglement and coherence of two coupled qubits each interacting with a local boson or fermion reservoir, based on the Bloch-Redfield master equation beyond the secular approximation. We find that there is non-vanishing steady-state coherence in the nonequilibrium scenario, which grows monotonically with the nonequilibrium condition quantified by the temperature difference or chemical potential difference of the two baths. The steady-state entanglement in general is a non-monotonic function of the nonequilibrium condition as well as the bath parameters in the equilibrium setting. We also find that weak inter-qubit coupling and high base temperature or chemical potential of the baths can strongly suppress the steady-state entanglement and coherence, regardless of the strength of the nonequilibrium condition. On the other hand, the energy detuning of the two qubits, when used in a compensatory way with the nonequilibrium condition, can lead to significant enhancement of the steady-state entanglement in some parameter regimes. In addition, the qubits typically have a stronger steady-state entanglement when coupled to fermion baths exchanging particle with the system than boson baths exchanging energy with the system under similar conditions. We also discussed the possible experimental realization of measuring the steady state entanglement and coherence for coupled qubits systems in nonequilibrium environments. These results offer some general guidelines for optimizing the steady-state entanglement and coherence in the coupled qubit system and may find potential applications in quantum information technology.

quant-ph↗

arXiv2019

Constants of de Bruijn-Newman type in analytic number theory and statistical physics

One formulation in 1859 of the Riemann Hypothesis (RH) was that the Fourier transform $H_f(z)$ of $f$ for $ z \in \mathbb{C}$ has only real zeros when $f(t)$ is a specific function $Φ(t)$. Pólya's 1920s approach to RH extended $H_f$ to $H_{f,λ}$, the Fourier transform of $e^{λt^2} f(t)$. We review developments of this approach to RH and related ones in statistical physics where $f(t)$ is replaced by a measure $d ρ(t)$. Pólya's work together with 1950 and 1976 results of de Bruijn and Newman, respectively, imply the existence of a finite constant $Λ_{DN} = Λ_{DN} (Φ)$ in $(-\infty, 1/2]$ such that $H_{Φ,λ}$ has only real zeros if and only if $λ\geq Λ_{DN}$; RH is then equivalent to $Λ_{DN} \leq 0$. Recent developments include the Rodgers and Tao proof of the 1976 conjecture that $Λ_{DN} \geq 0$ (that RH, if true, is only barely so) and the Polymath 15 project improving the $1/2$ upper bound to about $0.22$. We also present examples of $ρ$'s with differing $H_{ρ,λ}$ and $Λ_{DN} (ρ)$ behaviors; some of these are new and based on a recent weak convergence theorem of the authors.

math.PR↗

arXiv2019

Selective Sensor Fusion for Neural Visual-Inertial Odometry

Deep learning approaches for Visual-Inertial Odometry (VIO) have proven successful, but they rarely focus on incorporating robust fusion strategies for dealing with imperfect input sensory data. We propose a novel end-to-end selective sensor fusion framework for monocular VIO, which fuses monocular images and inertial measurements in order to estimate the trajectory whilst improving robustness to real-life issues, such as missing and corrupted data or bad sensor synchronization. In particular, we propose two fusion modalities based on different masking strategies: deterministic soft fusion and stochastic hard fusion, and we compare with previously proposed direct fusion baselines. During testing, the network is able to selectively process the features of the available sensor modalities and produce a trajectory at scale. We present a thorough investigation on the performances on three public autonomous driving, Micro Aerial Vehicle (MAV) and hand-held VIO datasets. The results demonstrate the effectiveness of the fusion strategies, which offer better performances compared to direct fusion, particularly in presence of corrupted data. In addition, we study the interpretability of the fusion networks by visualising the masking layers in different scenarios and with varying data corruption, revealing interesting correlations between the fusion networks and imperfect sensory input data.

cs.CV↗

arXiv2019

An Antenna Array Initial Condition Calibration Method for Integrated Optical Phased Array

This paper presents a modified rotating element electric field vector (modified REV) method to calibrate the antenna array initial condition of an optical phased array (OPA) device. The new method follows a similar sequential individual antenna phase calibration process while it modifies the algorithm to avoid possible π phase error in traditional REV when large initial phase distribution and finite optical power measurement accuracy present. We show that the method produces statistically more accurate and predictable calibration result which is highly desired in practice.

physics.app-ph↗

arXiv2019

Multi-Object Tracking with Multiple Cues and Switcher-Aware Classification

In this paper, we propose a unified Multi-Object Tracking (MOT) framework learning to make full use of long term and short term cues for handling complex cases in MOT scenes. Besides, for better association, we propose switcher-aware classification (SAC), which takes the potential identity-switch causer (switcher) into consideration. Specifically, the proposed framework includes a Single Object Tracking (SOT) sub-net to capture short term cues, a re-identification (ReID) sub-net to extract long term cues and a switcher-aware classifier to make matching decisions using extracted features from the main target and the switcher. Short term cues help to find false negatives, while long term cues avoid critical mistakes when occlusion happens, and the SAC learns to combine multiple cues in an effective way and improves robustness. The method is evaluated on the challenging MOT benchmarks and achieves the state-of-the-art results.

cs.CV↗

arXiv2019

Superconductivity in the WP single crystal

We report the discovery of superconductivity on high-quality single crystals of transition-metal pnictides WP grown by chemical vapor transport (CVT) method. Bulk superconductivity is observed at Tc = 0.84 K under ambient pressure by electrical resistivity and AC magnetic susceptibility measurements. The effects of magnetic field on the superconducting transitions are studied, leading to a large anisotropy parameter around 2 with the in-plane and out-of-plane upper critical fields of Hc2=172 Oe and Hc2\=85 Oe, respectively. Our finding demonstrates that WP is the first superconductor in 5d transition-metal at ambient pressure in MnP-type, which will help to search for new superconductors in transition-metal pnictides.

cond-mat.supr-con↗

arXiv2018

SiamRPN++: Evolution of Siamese Visual Tracking with Very Deep Networks

Siamese network based trackers formulate tracking as convolutional feature cross-correlation between target template and searching region. However, Siamese trackers still have accuracy gap compared with state-of-the-art algorithms and they cannot take advantage of feature from deep networks, such as ResNet-50 or deeper. In this work we prove the core reason comes from the lack of strict translation invariance. By comprehensive theoretical analysis and experimental validations, we break this restriction through a simple yet effective spatial aware sampling strategy and successfully train a ResNet-driven Siamese tracker with significant performance gain. Moreover, we propose a new model architecture to perform depth-wise and layer-wise aggregations, which not only further improves the accuracy but also reduces the model size. We conduct extensive ablation studies to demonstrate the effectiveness of the proposed tracker, which obtains currently the best results on four large tracking benchmarks, including OTB2015, VOT2018, UAV123, and LaSOT. Our model will be released to facilitate further studies based on this problem.

cs.CV↗

arXiv2018

A Review for Weighted MinHash Algorithms

Data similarity (or distance) computation is a fundamental research topic which underpins many high-level applications based on similarity measures in machine learning and data mining. However, in large-scale real-world scenarios, the exact similarity computation has become daunting due to "3V" nature (volume, velocity and variety) of big data. In such cases, the hashing techniques have been verified to efficiently conduct similarity estimation in terms of both theory and practice. Currently, MinHash is a popular technique for efficiently estimating the Jaccard similarity of binary sets and furthermore, weighted MinHash is generalized to estimate the generalized Jaccard similarity of weighted sets. This review focuses on categorizing and discussing the existing works of weighted MinHash algorithms. In this review, we mainly categorize the Weighted MinHash algorithms into quantization-based approaches, "active index"-based ones and others, and show the evolution and inherent connection of the weighted MinHash algorithms, from the integer weighted MinHash algorithms to real-valued weighted MinHash ones (particularly the Consistent Weighted Sampling scheme). Also, we have developed a python toolbox for the algorithms, and released it in our github. Based on the toolbox, we experimentally conduct a comprehensive comparative study of the standard MinHash algorithm and the weighted MinHash ones.

cs.DS↗