Searcharxiv⌕ Search

arXiv subjects

Wei Sun

Publications and source records attributed to Wei Sun.

At least 307 records · Page 17Linked to original sources

Explicit symplectic methods in black hole spacetimes

Many Hamiltonian problems in the Solar System are separable or separate into two analytically solvable parts, and thus give a great chance to the development and application of explicit symplectic integrators based on operator splitting and composing. However, such constructions cannot in general be available for curved spacetimes in general relativity and modified theories of gravity, because these curved spacetimes correspond to nonseparable Hamiltonians without the two part splits. Recently, several black hole spacetimes such as the Schwarzschild black hole were found to allow the construction of explicit symplectic integrators, since their corresponding Hamiltonians are separable into more than two explicitly integrable pieces. Although some other curved spacetimes including the Kerr black hole do not exist such multi part splits, their corresponding appropriate time transformation Hamiltonians do. In fact, the key problem for the obtainment of symplectic analytically integrable decomposition algorithms is how to split these Hamiltonians or time transformation Hamiltonians. Considering this idea, we develop explicit sympelcetic schemes in curved spacetimes. We introduce a class of spacetimes whose Hamiltonians are directly split into several explicitly integrable terms. For example, the Hamiltonian of rotating black ring has a 13 part split. We also present two sets of spacetimes whose appropriate time transformation Hamiltonians have the desirable splits. For instance, an 8 part split exists in a time-transformed Hamiltonian of Kerr-Newman solution with disformal parameter. In this way, the proposed symplectic splitting methods will be used widely for long-term integrations of orbits in most curved spacetimes we have known.

gr-qc↗

A Deep Learning based No-reference Quality Assessment Model for UGC Videos

Quality assessment for User Generated Content (UGC) videos plays an important role in ensuring the viewing experience of end-users. Previous UGC video quality assessment (VQA) studies either use the image recognition model or the image quality assessment (IQA) models to extract frame-level features of UGC videos for quality regression, which are regarded as the sub-optimal solutions because of the domain shifts between these tasks and the UGC VQA task. In this paper, we propose a very simple but effective UGC VQA model, which tries to address this problem by training an end-to-end spatial feature extraction network to directly learn the quality-aware spatial feature representation from raw pixels of the video frames. We also extract the motion features to measure the temporal-related distortions that the spatial features cannot model. The proposed model utilizes very sparse frames to extract spatial features and dense frames (i.e. the video chunk) with a very low spatial resolution to extract motion features, which thereby has low computational complexity. With the better quality-aware features, we only use the simple multilayer perception layer (MLP) network to regress them into the chunk-level quality scores, and then the temporal average pooling strategy is adopted to obtain the video-level quality score. We further introduce a multi-scale quality fusion strategy to solve the problem of VQA across different spatial resolutions, where the multi-scale weights are obtained from the contrast sensitivity function of the human visual system. The experimental results show that the proposed model achieves the best performance on five popular UGC VQA databases, which demonstrates the effectiveness of the proposed model. The code will be publicly available.

cs.CV↗

Using Sums-of-Squares to Prove Gaussian Product Inequalities

The long-standing Gaussian product inequality (GPI) conjecture states that $E [\prod_{j=1}^{n}X_j^{2m_j}]\geq\prod_{j=1}^{n}E[X_j^{2m_j}]$ for any centered Gaussian random vector $(X_1,\dots,X_n)$ and $m_1,\dots,m_n\in\mathbb{N}$. In this paper, we describe a computational algorithm involving sums-of-squares representations of multivariate polynomials that can be used to resolve the GPI conjecture. To exhibit the power of this novel method, we apply it to prove two new GPIs: $E[X_1^{2m_1}X_2^{6}X_3^{4}]\ge E[X_1^{2m_1}]E[X_2^{6}]E[X_3^{4}]$ and $E[X_1^{2m_1}X_2^{2}X_3^{2}X_4^{2}]\ge E[X_1^{2m_1}]E[X_2^{2}]E[X_3^{2}]E[X_4^{2}]$.

math.PR↗

A flexible and scalable single-level framework for OD matrix inference using multiple sources of transport information

This study proposes a flexible and scalable single-level framework for origin-destination matrix (ODM) inference using data from IoT (Internet of Things) and other sources. The framework allows the analyst to integrate information from multiple data sources, while controlling for differences in data quality across sources. We assess the effectiveness of the framework through a real-world experiment in Greater Adelaide (GA), Australia. We infer car OD flows within the region using four separate data sources: site-level traffic counts from loop detectors, vehicle trajectories recorded by roadside Bluetooth sensors, partial OD flows based on data from in-vehicle navigation systems, and journey-to-work data collected by the Australian Census. We compare our OD inferences with those from the current version of the Metropolitan Adelaide Strategic Transport Model (MASTEM), calibrated using data from traditional household travel surveys. We find remarkable consistency between our inferences and those from MASTEM, despite differences in input data and methodologies. For example, for the morning peak period, we predict the total number of trips made within GA to be equal to 556,000, while the corresponding prediction from MASTEM is equal to 484,000. The two predictions are within 20 per cent of each other. When we compare the spatial distribution of trips, in terms of origins and destinations, we find that our inferred OD matrix has an 86 per cent cosine similarity to the corresponding MASTEM matrix. In summary, our results show that the proposed framework can produce highly comparable ODM, trip production and trip attraction patterns to those inferred from traditional household travel survey-based transportation demand modelling methods.

physics.soc-ph↗

A No-reference Quality Assessment Metric for Point Cloud Based on Captured Video Sequences

Point cloud is one of the most widely used digital formats of 3D models, the visual quality of which is quite sensitive to distortions such as downsampling, noise, and compression. To tackle the challenge of point cloud quality assessment (PCQA) in scenarios where reference is not available, we propose a no-reference quality assessment metric for colored point cloud based on captured video sequences. Specifically, three video sequences are obtained by rotating the camera around the point cloud through three specific orbits. The video sequences not only contain the static views but also include the multi-frame temporal information, which greatly helps understand the human perception of the point clouds. Then we modify the ResNet3D as the feature extraction model to learn the correlation between the capture videos and corresponding subjective quality scores. The experimental results show that our method outperforms most of the state-of-the-art full-reference and no-reference PCQA metrics, which validates the effectiveness of the proposed method.

eess.IV↗

Moment ratio inequality of bivariate Gaussian distribution and three-dimensional Gaussian product inequality

We prove the three-dimensional Gaussian product inequality (GPI) $E[X_1^{2}X_2^{2m_2}X_3^{2m_3}]\ge E[X_1^{2}]E[X_2^{2m_2}]E[X_3^{2m_3}]$ for any centered Gaussian random vector $(X_1,X_2,X_3)$ and $m_2,m_3\in\mathbb{N}$. We discover a novel inequality for the moment ratio $\frac{|E[ X_2^{2m_2+1}X_3^{2m_3+1}]|}{E[ X_2^{2m_2}X_3^{2m_3}]}$, which implies the 3D-GPI. The interplay between computing and hard analysis plays a crucial role in the proofs.

math.PR↗

Multitask Balanced and Recalibrated Network for Medical Code Prediction

Human coders assign standardized medical codes to clinical documents generated during patients' hospitalization, which is error-prone and labor-intensive. Automated medical coding approaches have been developed using machine learning methods such as deep neural networks. Nevertheless, automated medical coding is still challenging because of the imbalanced class problem, complex code association, and noise in lengthy documents. To solve these issues, we propose a novel neural network called Multitask Balanced and Recalibrated Neural Network. Significantly, the multitask learning scheme shares the relationship knowledge between different code branches to capture the code association. A recalibrated aggregation module is developed by cascading convolutional blocks to extract high-level semantic features that mitigate the impact of noise in documents. Also, the cascaded structure of the recalibrated module can benefit the learning from lengthy notes. To solve the class imbalanced problem, we deploy the focal loss to redistribute the attention of low and high-frequency medical codes. Experimental results show that our proposed model outperforms competitive baselines on a real-world clinical dataset MIMIC-III.

cs.CL↗

A Distributionally Robust Resilience Enhancement Strategy for Distribution Networks Considering Decision-Dependent Contingencies

When performing the resilience enhancement for distribution networks, there are two obstacles to reliably model the uncertain contingencies: 1) decision-dependent uncertainty (DDU) due to various line hardening decisions, and 2) distributional ambiguity due to limited outage information during extreme weather events (EWEs). To address these two challenges, this paper develops scenario-wise decision-dependent ambiguity sets (SWDD-ASs), where the DDU and distributional ambiguity inherent in EWE-induced contingencies are simultaneously captured for each possible EWE scenario. Then, a two-stage trilevel decision-dependent distributionally robust resilient enhancement (DD-DRRE) model is formulated, whose outputs include the optimal line hardening, distributed generation (DG) allocation, and proactive network reconfiguration strategy under the worst-case distributions in SWDD-ASs. Subsequently, the DD-DRRE model is equivalently recast to a mixed-integer linear programming (MILP)-based master problem and multiple scenario-wise subproblems, facilitating the adoption of a customized column-and-constraint generation (C&CG) algorithm. Finally, case studies demonstrate a remarkable improvement in the out-of-sample performance of our model, compared to its prevailing stochastic and robust counterparts. Moreover, the potential values of incorporating the ambiguity and distributional information are quantitatively estimated, providing a useful reference for planners with different budgets and risk-aversion levels.

eess.SY↗

Towards Learning in Grey Spatiotemporal Systems: A Prophet to Non-consecutive Spatiotemporal Dynamics

Spatiotemporal forecasting is an imperative topic in data science due to its diverse and critical applications in smart cities. Existing works mostly perform consecutive predictions of following steps with observations completely and continuously obtained, where nearest observations can be exploited as key knowledge for instantaneous status estimation. However, the practical issues of early activity planning and sensor failures elicit a brand-new task, i.e., non-consecutive forecasting. In this paper, we define spatiotemporal learning systems with missing observation as Grey Spatiotemporal Systems (G2S) and propose a Factor-Decoupled learning framework for G2S (FDG2S), where the core idea is to hierarchically decouple multi-level factors and enable both flexible aggregations and disentangled uncertainty estimations. Firstly, to compensate for missing observations, a generic semantic-neighboring sequence sampling is devised, which selects representative sequences to capture both periodical regularity and instantaneous variations. Secondly, we turn the predictions of non-consecutive statuses into inferring statuses under expected combined exogenous factors. In particular, a factor-decoupled aggregation scheme is proposed to decouple factor-induced predictive intensity and region-wise proximity by two energy functions of conditional random field. To infer region-wise proximity under flexible factor-wise combinations and enable dynamic neighborhood aggregations, we further disentangle compounded influences of exogenous factors on region-wise proximity and learn to aggregate them. Given the inherent incompleteness and critical applications of G2S, a DisEntangled Uncertainty Quantification is put forward, to identify two types of uncertainty for reliability guarantees and model interpretations.

cs.LG↗

$T_{cc}^{+}(3875)$ relevant $DD^*$ scattering from $N_f=2$ lattice QCD

The $S$-wave $DD^*$ scattering in the isospin $I=0,1$ channels is studied in $N_f=2$ lattice QCD at $m_π\approx 350$ MeV. It is observed that the $DD^*$ interaction is repulsive in the $I=1$ channel when the $DD^*$ energy is near the $DD^*$ threshold. In contrast, the $DD^*$ interaction in the $I=0$ channel is definitely attractive in a wide range of the $DD^*$ energy. This is consistent with the isospin assignment $I=0$ for $T_{cc}^+(3875)$. By analyzing the components of the $DD^*$ correlation functions, it turns out that the quark diagram responsible for the different properties of $I=0,1$ $DD^*$ interactions can be understood as the charged $ρ$ meson exchange effect. This observation provides direct information on the internal dynamics of $T_{cc}^+(3875)$.

hep-lat↗

Domain-invariant Prototypes for Semantic Segmentation

Deep Learning has greatly advanced the performance of semantic segmentation, however, its success relies on the availability of large amounts of annotated data for training. Hence, many efforts have been devoted to domain adaptive semantic segmentation that focuses on transferring semantic knowledge from a labeled source domain to an unlabeled target domain. Existing self-training methods typically require multiple rounds of training, while another popular framework based on adversarial training is known to be sensitive to hyper-parameters. In this paper, we present an easy-to-train framework that learns domain-invariant prototypes for domain adaptive semantic segmentation. In particular, we show that domain adaptation shares a common character with few-shot learning in that both aim to recognize some types of unseen data with knowledge learned from large amounts of seen data. Thus, we propose a unified framework for domain adaptation and few-shot learning. The core idea is to use the class prototypes extracted from few-shot annotated target images to classify pixels of both source images and target images. Our method involves only one-stage training and does not need to be trained on large-scale un-annotated target images. Moreover, our method can be extended to variants of both domain adaptation and few-shot learning. Experiments on adapting GTA5-to-Cityscapes and SYNTHIA-to-Cityscapes show that our method achieves competitive performance to state-of-the-art.

cs.CV↗

Quantum phase with coexisting localized, extended, and critical zones

Conventionally a mobility edge (ME) marks a critical energy that separates two different transport zones where all states are extended and localized, respectively. Here we propose a novel quasiperiodic spin-orbit coupled lattice model with experimental feasibility to realize a fundamentally new quantum phase with three coexisting energy-dependent zones, i.e., the extended, critical and localized zones, and uncover the underlying generic mechanism for the occurrence of the new quantum phase. Accordingly, this phase exhibits new types of MEs which separate the extended states from critical ones, and the localized states from critical ones, respectively. We introduce the diagnostic quantities to characterize and distinguish the different zones, and show that the predicted phase can be detected by measuring the fractal dimension or conductivities. The experimental realization is also proposed and studied. This work extends the concept of ME and enriches the quantum phases in disordered systems, which sheds light on searching for new localization and critical phenomena with novel transport and thermoelectric effects.

cond-mat.dis-nn↗

Weak Random Periodic Solutions of Random Dynamical Systems

We first introduce the concept of weak random periodic solutions of random dynamical systems. Then, we discuss the existence of such periodic solutions. Further, we introduce the definition of weak random periodic measures and study their relationship with weak random periodic solutions. In particular, we establish the existence of invariant measures of random dynamical systems by virtue of their weak random periodic solutions. Finally, we use concrete examples to illustrate the weak random periodic phenomena of dynamical systems induced by random and stochastic differential equations.

math.DS↗

Visual-Assisted Sound Source Depth Estimation in the Wild

Depth estimation enables a wide variety of 3D applications, such as robotics, autonomous driving, and virtual reality. Despite significant work in this area, it remains open how to enable accurate, low-cost, high-resolution, and large-range depth estimation. Inspired by the flash-to-bang phenomenon (i.e. hearing the thunder after seeing the lightning), this paper develops FBDepth, the first audio-visual depth estimation framework. It takes the difference between the time-of-flight (ToF) of the light and the sound to infer the sound source depth. FBDepth is the first to incorporate video and audio with both semantic features and spatial hints for range estimation. It first aligns correspondence between the video track and audio track to locate the target object and target sound in a coarse granularity. Based on the observation of moving objects' trajectories, FBDepth proposes to estimate the intersection of optical flow before and after the sound production to locate video events in time. FBDepth feeds the estimated timestamp of the video event and the audio clip for the final depth estimation. We use a mobile phone to collect 3000+ video clips with 20 different objects at up to $50m$. FBDepth decreases the Absolute Relative error (AbsRel) by 55\% compared to RGB-based methods.

cs.SD↗

Data-driven discovery of quasi-disordered mechanical metamaterials failed progressively

Natural cellular materials, such as honeycombs, woods, foams, trabecular bones, plant parenchyma, and sponges, may benefit from the disorderliness within their internal microstructures to achieve damage tolerant behaviours. Inspired by this, we have created quasi-disordered truss metamaterials (QTMs) via introducing spatial coordinate perturbations or strut thickness variations to the perfect, periodic truss lattices. Numerical studies have suggested that the QTMs can exhibit either ductile, damage tolerant behaviours or sudden, catastrophic failure mode, depending on the distribution of the introduced disorderliness. A data-driven approach has been developed, combining deep-learning and global optimization algorithms, to tune the distribution of the disorderliness to achieve the damage tolerant QTM designs. A case study on the QTMs created from a periodic Face Centred Cubic (FCC) lattice has demonstrated that the optimised QTMs can achieve up to 100% increase in ductility at the expense of less than 5% stiffness and less than 10% tensile strength. Our results suggest a novel design pathway for architected materials to improve damage tolerance.

cond-mat.dis-nn↗

Constrained Prescriptive Trees via Column Generation

With the abundance of available data, many enterprises seek to implement data-driven prescriptive analytics to help them make informed decisions. These prescriptive policies need to satisfy operational constraints, and proactively eliminate rule conflicts, both of which are ubiquitous in practice. It is also desirable for them to be simple and interpretable, so they can be easily verified and implemented. Existing approaches from the literature center around constructing variants of prescriptive decision trees to generate interpretable policies. However, none of the existing methods are able to handle constraints. In this paper, we propose a scalable method that solves the constrained prescriptive policy generation problem. We introduce a novel path-based mixed-integer program (MIP) formulation which identifies a (near) optimal policy efficiently via column generation. The policy generated can be represented as a multiway-split tree which is more interpretable and informative than a binary-split tree due to its shorter rules. We demonstrate the efficacy of our method with extensive experiments on both synthetic and real datasets.

math.OC↗

Spatial Aware Multi-Task Learning Based Speech Separation

During the Covid, online meetings have become an indispensable part of our lives. This trend is likely to continue due to their convenience and broad reach. However, background noise from other family members, roommates, office-mates not only degrades the voice quality but also raises serious privacy issues. In this paper, we develop a novel system, called Spatial Aware Multi-task learning-based Separation (SAMS), to extract audio signals from the target user during teleconferencing. Our solution consists of three novel components: (i) generating fine-grained location embeddings from the user's voice and inaudible tracking sound, which contains the user's position and rich multipath information, (ii) developing a source separation neural network using multi-task learning to jointly optimize source separation and location, and (iii) significantly speeding up inference to provide a real-time guarantee. Our testbed experiments demonstrate the effectiveness of our approach

cs.SD↗

Smooth Solutions to Asymptotic Plateau Type Problem in Hyperbolic Space

We investigate on the existence of smooth complete hypersurface with prescribed Weingarten curvature and asymptotic boundary at infinity in hyperbolic space under the assumption that there exists an asymptotic subsolution. We give an affirmative answer for the case $k = n$ when the asymptotic boundary $Γ$ bounds a uniformly convex domain, and for $k < n$ when $Γ$ bounds a disk, utilizing Pogorelov type interior second order estimate. Our result complements our previous work \cite{Sui2019, Sui-Sun}, and generalizes the asymptotic Plateau type problem to non-constant prescribed curvature case.

math.DG↗