arXiv subjects

Jian Ma

Publications and source records attributed to Jian Ma.

At least 109 records · Page 6Linked to original sources

arXiv2020

A Gradual Takeover Strategy of the Active Safety System

A gradual takeover strategy is proposed, in which the dynamic driving privilege assignment in real-time and the driving privilege gradual handover are realized. Firstly, the driving privilege assignment based on the risk level is achieved. The naturalistic driving data is applied to study the driver behavior during danger. TTC (time to collision) is defined as an obvious risk measure, whereas the time before the host vehicle has to brake assuming that the target vehicle is braking is defined as the potential risk measure, i.e. the time margin (TM). A risk assessment algorithm is proposed based on the obvious risk and potential risk. Secondly, the driving privilege gradual handover is realized. The non-cooperative MPC (model predictive control) is employed to resolve the conflicts between the driver and active safety system. The naturalistic driving data are applied to verify the effectiveness of the risk assessment algorithm, and the risk assessment algorithm performs better than TTC in the ROC (receiver operating characteristic). It is identified that the Nash equilibrium of the non-cooperative MPC can be achieved by using a non-iterative method. The driving privilege gradual handover is realized by using the confidence matrixes updating. The simulation verification shows that the gradual takeover strategy can achieve the driving privilege gradual handover between the driver and active safety system.

eess.SY↗

arXiv2020

Variable Selection with Copula Entropy

Variable selection is of significant importance for classification and regression tasks in machine learning and statistical applications where both predictability and explainability are needed. In this paper, a Copula Entropy (CE) based method for variable selection which use CE based ranks to select variables is proposed. The method is both model-free and tuning-free. Comparison experiments between the proposed method and traditional variable selection methods, such as Distance Correlation, Hilbert-Schmidt Independence Criterion, Stepwise Selection, regularized generalized linear models and Adaptive LASSO, were conducted on the UCI heart disease data. Experimental results show that CE based method can select the `right' variables out more effectively and derive better interpretable results than traditional methods do without sacrificing accuracy performance. It is believed that CE based variable selection can help to build more explainable models.

cs.LG↗

arXiv2020

Predicting TUG score from gait characteristics with video analysis and machine learning

Fall is a leading cause of death which suffers the elderly and society. Timed Up and Go (TUG) test is a common tool for fall risk assessment. In this paper, we propose a method for predicting TUG score from gait characteristics extracted from video with computer vision and machine learning technologies. First, 3D pose is estimated from video captured with 2D and 3D cameras during human motion and then a group of gait characteristics are computed from 3D pose series. After that, copula entropy is used to select those characteristics which are mostly associated with TUG score. Finally, the selected characteristics are fed into the predictive models to predict TUG score. Experiments on real world data demonstrated the effectiveness of the proposed method. As a byproduct, the associations between TUG score and several gait characteristics are discovered, which laid the scientific foundation of the proposed method and make the predictive models such built interpretable to clinical users.

cs.CV↗

arXiv2020

Discovering Association with Copula Entropy

Discovering associations is of central importance in scientific practices. Currently, most researches consider only linear association measured by correlation coefficient, which has its theoretical limitations. In this paper, we propose a new method for discovering association with copula entropy -- a universal applicable association measure for not only linear cases, but nonlinear cases. The advantage of the method based on copula entropy over traditional method is demonstrated on the NHANES data by discovering more biomedical meaningful associations.

cs.LG↗

arXiv2020

Graph Representation Learning for Merchant Incentive Optimization in Mobile Payment Marketing

Mobile payment such as Alipay has been widely used in our daily lives. To further promote the mobile payment activities, it is important to run marketing campaigns under a limited budget by providing incentives such as coupons, commissions to merchants. As a result, incentive optimization is the key to maximizing the commercial objective of the marketing campaign. With the analyses of online experiments, we found that the transaction network can subtly describe the similarity of merchants' responses to different incentives, which is of great use in the incentive optimization problem. In this paper, we present a graph representation learning method atop of transaction networks for merchant incentive optimization in mobile payment marketing. With limited samples collected from online experiments, our end-to-end method first learns merchant representations based on an attributed transaction networks, then effectively models the correlations between the commercial objectives each merchant may achieve and the incentives under varying treatments. Thus we are able to model the sensitivity to incentive for each merchant, and spend the most budgets on those merchants that show strong sensitivities in the marketing campaign. Extensive offline and online experimental results at Alipay demonstrate the effectiveness of our proposed approach.

cs.SI↗

arXiv2019

Hyper-SAGNN: a self-attention based graph neural network for hypergraphs

Graph representation learning for hypergraphs can be used to extract patterns among higher-order interactions that are critically important in many real world problems. Current approaches designed for hypergraphs, however, are unable to handle different types of hypergraphs and are typically not generic for various learning tasks. Indeed, models that can predict variable-sized heterogeneous hyperedges have not been available. Here we develop a new self-attention based graph neural network called Hyper-SAGNN applicable to homogeneous and heterogeneous hypergraphs with variable hyperedge sizes. We perform extensive evaluations on multiple datasets, including four benchmark network datasets and two single-cell Hi-C datasets in genomics. We demonstrate that Hyper-SAGNN significantly outperforms the state-of-the-art methods on traditional tasks while also achieving great performance on a new task called outsider identification. Hyper-SAGNN will be useful for graph representation learning to uncover complex higher-order interactions in different applications.

cs.LG↗

arXiv2019

Dependence Structure Estimation via Copula

Dependence strucuture estimation is one of the important problems in machine learning domain and has many applications in different scientific areas. In this paper, a theoretical framework for such estimation based on copula and copula entropy -- the probabilistic theory of representation and measurement of statistical dependence, is proposed. Graphical models are considered as a special case of the copula framework. A method of the framework for estimating maximum spanning copula is proposed. Due to copula, the method is irrelevant to the properties of individual variables, insensitive to outlier and able to deal with non-Gaussianity. Experiments on both simulated data and real dataset demonstrated the effectiveness of the proposed method.

cs.LG↗

arXiv2018

Rotation Equivariance and Invariance in Convolutional Neural Networks

Performance of neural networks can be significantly improved by encoding known invariance for particular tasks. Many image classification tasks, such as those related to cellular imaging, exhibit invariance to rotation. We present a novel scheme using the magnitude response of the 2D-discrete-Fourier transform (2D-DFT) to encode rotational invariance in neural networks, along with a new, efficient convolutional scheme for encoding rotational equivariance throughout convolutional layers. We implemented this scheme for several image classification tasks and demonstrated improved performance, in terms of classification accuracy, time required to train the model, and robustness to hyperparameter selection, over a standard CNN and another state-of-the-art method.

stat.ML↗

arXiv2018

On the Reconstruction Risk of Convolutional Sparse Dictionary Learning

Sparse dictionary learning (SDL) has become a popular method for adaptively identifying parsimonious representations of a dataset, a fundamental problem in machine learning and signal processing. While most work on SDL assumes a training dataset of independent and identically distributed samples, a variant known as convolutional sparse dictionary learning (CSDL) relaxes this assumption, allowing more general sequential data sources, such as time series or other dependent data. Although recent work has explored the statistical properties of classical SDL, the statistical properties of CSDL remain unstudied. This paper begins to study this by identifying the minimax convergence rate of CSDL in terms of reconstruction risk, by both upper bounding the risk of an established CSDL estimator and proving a matching information-theoretic lower bound. Our results indicate that consistency in reconstruction risk is possible precisely in the `ultra-sparse' setting, in which the sparsity (i.e., the number of feature occurrences) is in $o(N)$ in terms of the length N of the training sequence. Notably, our results make very weak assumptions, allowing arbitrary dictionaries and dependent measurement noise. Finally, we verify our theoretical results with numerical experiments on synthetic data.

math.ST↗

arXiv2018

Correlating Cellular Features with Gene Expression using CCA

To understand the biology of cancer, joint analysis of multiple data modalities, including imaging and genomics, is crucial. The involved nature of gene-microenvironment interactions necessitates the use of algorithms which treat both data types equally. We propose the use of canonical correlation analysis (CCA) and a sparse variant as a preliminary discovery tool for identifying connections across modalities, specifically between gene expression and features describing cell and nucleus shape, texture, and stain intensity in histopathological images. Applied to 615 breast cancer samples from The Cancer Genome Atlas, CCA revealed significant correlation of several image features with expression of PAM50 genes, known to be linked to outcome, while Sparse CCA revealed associations with enrichment of pathways implicated in cancer without leveraging prior biological understanding. These findings affirm the utility of CCA for joint phenotype-genotype analysis of cancer.

eess.SP↗

arXiv2017

Ionic Behavior in Highly Concentrated Aqueous Solutions Nanoconfined between Discretely Charged Silicon Surfaces

Through molecular dynamics simulations considering thermal vibration of surface atoms, ionic behaviors in concentrated NaCl solutions confined between discretely charged silicon surfaces have been investigated. The electric double layer structure was found sensitive to the density and distribution of surface charges. Due to the surface charge discreteness, slight charge inversion appeared which depended on the surface charge density, bulk concentration and confinement. In the nanoconfined NaCl solutions differently concentrated from 0.2 M to 4.0 M, the locations of accumulation layers for Na+ and Cl- ions kept stable, but their peak values increased. The higher the concentration was, the more obvious charge inversion appeared. In 4.0 M NaCl solution, Na+ and Cl- ions show obvious alternating layered distributions which may be corresponding to the solidification found in experiments. By changing surface separation, the confinement had a large effect on ionic distributions. As both surfaces approached each other, many ions and water molecules were squeezed out of the confined space. Two adjacent layers in ion or water distribution profiles can be forced to closer to each other and merge together. From ionic hydration analysis, the coordination number of Na+ ions in highly-confined space was much lower than that in the bulk.

physics.chem-ph↗

arXiv2017

Speeding Up Latent Variable Gaussian Graphical Model Estimation via Nonconvex Optimizations

We study the estimation of the latent variable Gaussian graphical model (LVGGM), where the precision matrix is the superposition of a sparse matrix and a low-rank matrix. In order to speed up the estimation of the sparse plus low-rank components, we propose a sparsity constrained maximum likelihood estimator based on matrix factorization, and an efficient alternating gradient descent algorithm with hard thresholding to solve it. Our algorithm is orders of magnitude faster than the convex relaxation based methods for LVGGM. In addition, we prove that our algorithm is guaranteed to linearly converge to the unknown sparse and low-rank components up to the optimal statistical precision. Experiments on both synthetic and genomic data demonstrate the superiority of our algorithm over the state-of-the-art algorithms and corroborate our theory.

stat.ML↗

arXiv2016

Design considerations of high-performance InGaAs/InP single-photon avalanche diodes for quantum key distribution

InGaAs/InP single-photon avalanche diodes (SPADs) are widely used in practical applications requiring near-infrared photon counting such as quantum key distribution (QKD). Photon detection efficiency and dark count rate are the intrinsic parameters of InGaAs/InP SPADs, due to the fact that their performances cannot be improved using different quenching electronics given the same operation conditions. After modeling these parameters and developing a simulation platform for InGaAs/InP SPADs, we investigate the semiconductor structure design and optimization. The parameters of photon detection efficiency and dark count rate highly depend on the variables of absorption layer thickness, multiplication layer thickness, excess bias voltage and temperature. By evaluating the decoy-state QKD performance, the variables for SPAD design and operation can be globally optimized. Such optimization from the perspective of specific applications can provide an effective approach to design high-performance InGaAs/InP SPADs.

quant-ph↗

arXiv2016

Pedestrian dynamics in single-file movement of crowd with different age compositions

An aging population is bringing new challenges to the management of escape routes and facility design in many countries. This paper investigates pedestrian movement properties of crowd with different age compositions. Three pedestrian groups are considered: young student group, old people group and mixed group. It is found that traffic jams occur more frequently in mixed group due to the great differences of mobilities and self-adaptive abilities among pedestrians. The jams propagate backward with a velocity 0.4 m/s for global density around 1.75 m-1 and 0.3 m/s for higher than 2.3 m-1. The fundamental diagrams of the three groups are obviously different from each other and cannot be unified into one diagram by direct non-dimensionalization. Unlike previous studies, three linear regimes in mixed group but only two regimes in young student group are observed in the headway-velocity relation, which is also verified in the fundamental diagram. Different ages and mobilities of pedestrians in a crowd cause the heterogeneity of system and influence the properties of pedestrian dynamics significantly. It indicates that the density is not the only factor leading to jams in pedestrian traffic. The composition of crowd has to be considered in understanding pedestrian dynamics and facility design.

physics.soc-ph↗

arXiv2016

A new correlation clustering method for cancer mutation analysis

Cancer genomes exhibit a large number of different alterations that affect many genes in a diverse manner. It is widely believed that these alterations follow combinatorial patterns that have a strong connection with the underlying molecular interaction networks and functional pathways. A better understanding of the generative mechanisms behind the mutation rules and their influence on gene communities is of great importance for the process of driver mutations discovery and for identification of network modules related to cancer development and progression. We developed a new method for cancer mutation pattern analysis based on a constrained form of correlation clustering. Correlation clustering is an agnostic learning method that can be used for general community detection problems in which the number of communities or their structure is not known beforehand. The resulting algorithm, named $C^3$, leverages mutual exclusivity of mutations, patient coverage, and driver network concentration principles; it accepts as its input a user determined combination of heterogeneous patient data, such as that available from TCGA (including mutation, copy number, and gene expression information), and creates a large number of clusters containing mutually exclusive mutated genes in a particular type of cancer. The cluster sizes may be required to obey some useful soft size constraints, without impacting the computational complexity of the algorithm. To test $C^3$, we performed a detailed analysis on TCGA breast cancer and glioblastoma data and showed that our algorithm outperforms the state-of-the-art CoMEt method in terms of discovering mutually exclusive gene modules and identifying driver genes. Our $C^3$ method represents a unique tool for efficient and reliable identification of mutation patterns and driver pathways in large-scale cancer genomics studies.

cs.LG↗

arXiv2015

Discrete Element Crowd Model for Pedestrian Evacuation Through an Exit

A series of accidents caused by crowd within the last decades evoked a lot of scientific interest in modeling the movement of pedestrian crowds. Based on discrete element method, a granular dynamic model, in which human body is simplified as self-driven sphere, is proposed to simulate the characteristics of crowd flow through an exit. In this model, the repulsive force among people is considered to have an anisotropic feature, and the physical contact force due to body deformation is quantified by the Hertz contact model. The movement of human body is simulated by applying the second Newton's law. The crowd flow through an exit at different desired velocities is studied and simulation results indicated that crowd flow exhibits three distinct states, i.e., smooth state, transition state and phase separation state. In the simulation, clogging phenomenon occurs more easily when the velocity of desired is high and the exit may as a result be totally blocked at a desired velocity of 1.6m/s or above, leading to faster-to-frozen effect.

physics.soc-ph↗

arXiv2015

Modeling pedestrian evacuation movement in a swaying ship

With the advance in living standard, cruise travel has been rapidly expanding around the world in recent years. The transportation of passengers in water has also made a rapid development. It is expected that ships will be more and more widely used. Unfortunately, ship disasters occurred in these years caused serious losses. It raised the concern on effectiveness of passenger evacuation on ships. The present study thus focuses on pedestrian evacuation features on ships. On ships, passenger movements are affected by the periodical water motion and thus are quite different from the characteristic when walking on static horizontal floor. Taking into consideration of this special feature, an agent-based pedestrian model is formulized and the effect of ship swaying on pedestrian evacuation efficiency is investigated. Results indicated that the proposed model can be used to quantify the special evacuation process on ships.

physics.soc-ph↗

arXiv2015

DNA-Based Storage: Trends and Methods

We provide an overview of current approaches to DNA-based storage system design and accompanying synthesis, sequencing and editing methods. We also introduce and analyze a suite of new constrained coding schemes for both archival and random access DNA storage channels. The mathematical basis of our work is the construction and design of sequences over discrete alphabets that avoid pre-specified address patterns, have balanced base content, and exhibit other relevant substring constraints. These schemes adapt the stored signals to the DNA medium and thereby reduce the inherent error-rate of the system.

cs.ET↗