arXiv subjects

Yu Zheng

Publications and source records attributed to Yu Zheng.

266 records · Page 15Linked to original sources

arXiv2017

Deep Spatio-Temporal Residual Networks for Citywide Crowd Flows Prediction

Forecasting the flow of crowds is of great importance to traffic management and public safety, yet a very challenging task affected by many complex factors, such as inter-region traffic, events and weather. In this paper, we propose a deep-learning-based approach, called ST-ResNet, to collectively forecast the in-flow and out-flow of crowds in each and every region through a city. We design an end-to-end structure of ST-ResNet based on unique properties of spatio-temporal data. More specifically, we employ the framework of the residual neural networks to model the temporal closeness, period, and trend properties of the crowd traffic, respectively. For each property, we design a branch of residual convolutional units, each of which models the spatial properties of the crowd traffic. ST-ResNet learns to dynamically aggregate the output of the three residual neural networks based on data, assigning different weights to different branches and regions. The aggregation is further combined with external factors, such as weather and day of the week, to predict the final traffic of crowds in each and every region. We evaluate ST-ResNet based on two types of crowd flows in Beijing and NYC, finding that its performance exceeds six well-know methods.

cs.AI↗

arXiv2017

Predicting Citywide Crowd Flows Using Deep Spatio-Temporal Residual Networks

Forecasting the flow of crowds is of great importance to traffic management and public safety, and very challenging as it is affected by many complex factors, including spatial dependencies (nearby and distant), temporal dependencies (closeness, period, trend), and external conditions (e.g., weather and events). We propose a deep-learning-based approach, called ST-ResNet, to collectively forecast two types of crowd flows (i.e. inflow and outflow) in each and every region of a city. We design an end-to-end structure of ST-ResNet based on unique properties of spatio-temporal data. More specifically, we employ the residual neural network framework to model the temporal closeness, period, and trend properties of crowd traffic. For each property, we design a branch of residual convolutional units, each of which models the spatial properties of crowd traffic. ST-ResNet learns to dynamically aggregate the output of the three residual neural networks based on data, assigning different weights to different branches and regions. The aggregation is further combined with external factors, such as weather and day of the week, to predict the final traffic of crowds in each and every region. We have developed a real-time system based on Microsoft Azure Cloud, called UrbanFlow, providing the crowd flow monitoring and forecasting in Guiyang City of China. In addition, we present an extensive experimental evaluation using two types of crowd flows in Beijing and New York City (NYC), where ST-ResNet outperforms nine well-known baselines.

cs.AI↗

arXiv2016

Predicting Urban Water Quality with Ubiquitous Data

Urban water quality is of great importance to our daily lives. Prediction of urban water quality help control water pollution and protect human health. However, predicting the urban water quality is a challenging task since the water quality varies in urban spaces non-linearly and depends on multiple factors, such as meteorology, water usage patterns, and land uses. In this work, we forecast the water quality of a station over the next few hours from a data-driven perspective, using the water quality data and water hydraulic data reported by existing monitor stations and a variety of data sources we observed in the city, such as meteorology, pipe networks, structure of road networks, and point of interests (POIs). First, we identify the influential factors that affect the urban water quality via extensive experiments. Second, we present a multi-task multi-view learning method to fuse those multiple datasets from different domains into an unified learning model. We evaluate our method with real-world datasets, and the extensive experiments verify the advantages of our method over other baselines and demonstrate the effectiveness of our approach.

cs.CY↗

arXiv2016

On a conjecture of Chen-Guo-Wang

Towards confirming Sun's conjecture on the strict log-concavity of combinatorial sequence involving the n$th$ Bernoulli number, Chen, Guo and Wang proposed a conjecture about the log-concavity of the function $θ(x)=\sqrt[x]{2ζ(x)Γ(x+1)}$ for $x\in (6,\infty)$, where $ζ(x)$ is the Riemann zeta function and $Γ(x)$ is the Gamma function. In this paper, we first prove this conjecture along the spirit of Zhu's previous work. Second, we extend Chen et al.'s conjecture in the sense of almost infinite log-monotonicity of combinatorial sequences, which was also introduced by Chen et al. Furthermore, by using an analogue criterion to the one of Chen, Guo and Wang, we deduce the almost infinite log-monotonicity of the sequences $\frac{1}{\sqrt[n]{|B_{2n}|}}$, $T_n$ and $\frac{1}{\sqrt[n]{T_n}}$, where $B_{2n}$ and $T_{n}$ are the $2n$th Bernoulli number and the $n$th tangent number, respectively. These results can be seen as extensions of some solved conjectures of Sun.

math.CA↗

arXiv2014

Improved Deterministic N-To-One Joint Remote Preparation of an Arbitrary Qubit via EPR Pairs

Recently, Bich et al. (Int. J. Theor. Phys. 51: 2272, 2012) proposed two deterministic joint remote state preparation (JRSP) protocols of an arbitrary single-qubit state: one is for two preparers to remotely prepare for a receiver by using two Einstein-Podolsky-Rosen (ERP) pairs; the other is its generalized form in the case of arbitrary N>2 preparers via N ERP pairs. In this paper, Through reviewing and analyzing Bich et al.'s second protocols with N>2 preparers, we find that the success probability P_{suc}=1/4 < 1. In order to solve the problem, we firstly constructed two sets of projective measurement bases: the real-coefficient basis and the complex-coefficient one, and further proposed an improved deterministic N-to-one JRSP protocol for an arbitrary single-qubit state with unit success probability (i.e, P_{suc}=1). Morever, our protocol is also flexible and convenient, and it can be used in a practical network.

quant-ph↗

arXiv2014

Cryptanalysis and improvement of the quantum private comparison protocol based on Bell entangled states

Recently, Liu et al. [Commun. Theor. Phys. 57, 583, 2012] proposed a quantum private comparison protocol based on entanglement swapping of Bell states, which aims to securely compare the equality of two participants' information with the help of a semi-honest third party (TP). However, this study points out there is a fatal loophole in this protocol, i.e., TP can obtain all of the two participants secret inputs without being detected through making a specific Bell-basis measurement. To fix the problem, a simple solution, which uses one-time eavesdropper checking with decoy photons instead of twice eavesdropper checking with Bell states, is demonstrated. Compared with the original protocol, it also reduces the Bell states consumption and simplifies the steps in the protocol.

quant-ph↗

arXiv2014

PRESS: A Novel Framework of Trajectory Compression in Road Networks

Location data becomes more and more important. In this paper, we focus on the trajectory data, and propose a new framework, namely PRESS (Paralleled Road-Network-Based Trajectory Compression), to effectively compress trajectory data under road network constraints. Different from existing work, PRESS proposes a novel representation for trajectories to separate the spatial representation of a trajectory from the temporal representation, and proposes a Hybrid Spatial Compression (HSC) algorithm and error Bounded Temporal Compression (BTC) algorithm to compress the spatial and temporal information of trajectories respectively. PRESS also supports common spatial-temporal queries without fully decompressing the data. Through an extensive experimental study on real trajectory dataset, PRESS significantly outperforms existing approaches in terms of saving storage cost of trajectory data with bounded errors.

cs.DB↗

arXiv2012

Inferring Taxi Status Using GPS Trajectories

In this paper, we infer the statuses of a taxi, consisting of occupied, non-occupied and parked, in terms of its GPS trajectory. The status information can enable urban computing for improving a city's transportation systems and land use planning. In our solution, we first identify and extract a set of effective features incorporating the knowledge of a single trajectory, historical trajectories and geographic data like road network. Second, a parking status detection algorithm is devised to find parking places (from a given trajectory), dividing a trajectory into segments (i.e., sub-trajectories). Third, we propose a two-phase inference model to learn the status (occupied or non-occupied) of each point from a taxi segment. This model first uses the identified features to train a local probabilistic classifier and then carries out a Hidden Semi-Markov Model (HSMM) for globally considering long term travel patterns. We evaluated our method with a large-scale real-world trajectory dataset generated by 600 taxis, showing the advantages of our method over baselines.

cs.AI↗

arXiv2012

Reconciliation of Gene and Species Trees With Polytomies

Motivation: Millions of genes in the modern species belong to only thousands of `gene families'. A gene family includes instances of the same gene in different species (orthologs) and duplicate genes in the same species (paralogs). Genes are gained and lost during evolution. With advances in sequencing technology, researchers are able to investigate the important roles of gene duplications and losses in adaptive evolution. Because of gene complex evolution, ortholog identification is a basic but difficult task in comparative genomics. A key method for the task is to use an explicit model of the evolutionary history of the genes being studied, called the gene (family) tree. It compares the gene tree with the evolutionary history of the species in which the genes reside, called the species tree, using the procedure known as tree reconciliation. Reconciling binary gene and specific trees is simple. However, both gene and species trees may be non-binary in practice and thus tree reconciliation presents challenging problems. Here, non-binary gene and species tree reconciliation is studied in a binary refinement model. Results: The problem of reconciling arbitrary gene and species trees is proved NP-hard even for the duplication cost. We then present the first efficient method for reconciling a non-binary gene tree and a non-binary species tree. It attempts to find binary refinements of the given gene and species trees that minimize reconciliation cost. Our algorithms have been implemented into a software to support quick automated analysis of large data sets. Availability: The program, together with the source code, is available at its online server http://phylotoo.appspot.com.

q-bio.PE↗

arXiv2011

A Note on Conformal Ricci Flow

In this note we study conformal Ricci flow introduced by Arthur Fischer. We use DeTurck's trick to rewrite conformal Ricci flow as a strong parabolic-elliptic partial differential equations. Then we prove short time existences for conformal Ricci flow on compact manifolds as well as on asymptotically flat manifolds. We show that Yamabe constant is monotonically increasing along conformal Ricci flow on compact manifolds. We also show that conformal Ricci flow is the gradient flow for the ADM mass on asymptotically flat manifolds.

math.DG↗

arXiv2011

An Interpolating Curvature Condition Preserved By Ricci Flow

In this paper, we firstly establish an Interpolating curvature invariance between the well known nonnegative and 2-non-negative curvature invariant along the Ricci flow. Then a related strong maximum principle for the $(λ_1, λ_2)$-nonnegativity is also derived along Ricci flow. Based on these, finally we obtain a rigidity property of manifolds with $(λ_1,λ_2)$-nonnegative curvature operators.

math.DG↗

arXiv2011

Contracting convex hypersurfaces by curvature

We consider compact convex hypersurfaces contracting by functions of their curvature. Under the mean curvature flow, uniformly convex smooth initial hypersurfaces evolve to remain smooth and uniformly convex, and contract to points after finite time. The same holds if the initial data is only weakly convex or non-smooth, and the limiting shape at the final time is spherical. We provide a surprisingly large family of flows for which such results fail, by a variety of mechanisms: Uniformly convex hypersurfaces may become non-convex, and smooth ones may develop curvature singularities; even where this does not occur, non-uniformly convex regions and singular parts in the initial hypersurface may persist, including flat sides, ridges of infinite curvature, or `cylindrical' regions where some of the principal curvatures vanish; such cylindrical regions may persist even if the speed is positive, and in such cases the hypersurface may even collapse to a line segment or higher-dimensional disc rather than to a point. We provide sufficient conditions for these various disasters to occur, and by avoiding these arrive at a class of flows for which arbitrary weakly convex initial hypersurfaces immediately become smooth and uniformly convex and contract to points.

math.AP↗

arXiv2010

Relating diameter and mean curvature for Riemannian submanifolds

Given an $m$-dimensional closed connected Riemannian manifold $M$ smoothly isometrically immersed in an $n$-dimensional Riemannian manifold $N$, we estimate the diameter of $M$ in terms of its mean curvature field integral under some geometric restrictions, and therefore generalize a recent work of Topping in the Euclidean case (Comment. Math. Helv., 83 (2008), 539--546).

math.DG↗

arXiv2010

First eigenvalue of the $p$-Laplace operator along the Ricci flow

In this paper, we mainly investigate continuity, monotonicity and differentiability for the first eigenvalue of the $p$-Laplace operator along the Ricci flow on closed manifolds. We show that the first $p$-eigenvalue is strictly increasing and differentiable almost everywhere along the Ricci flow under some curvature assumptions. In particular, for an orientable closed surface, we construct various monotonic quantities and prove that the first $p$-eigenvalue is differentiable almost everywhere along the Ricci flow without any curvature assumption, and therefore derive a $p$-eigenvalue comparison-type theorem when its Euler characteristic is negative.

math.DG↗