arXiv subjects

Qing Xiao

Publications and source records attributed to Qing Xiao.

43 records · Page 3Linked to original sources

arXiv2024

Conditional Score-Based Diffusion Model for Cortical Thickness Trajectory Prediction

Alzheimer's Disease (AD) is a neurodegenerative condition characterized by diverse progression rates among individuals, with changes in cortical thickness (CTh) closely linked to its progression. Accurately forecasting CTh trajectories can significantly enhance early diagnosis and intervention strategies, providing timely care. However, the longitudinal data essential for these studies often suffer from temporal sparsity and incompleteness, presenting substantial challenges in modeling the disease's progression accurately. Existing methods are limited, focusing primarily on datasets without missing entries or requiring predefined assumptions about CTh progression. To overcome these obstacles, we propose a conditional score-based diffusion model specifically designed to generate CTh trajectories with the given baseline information, such as age, sex, and initial diagnosis. Our conditional diffusion model utilizes all available data during the training phase to make predictions based solely on baseline information during inference without needing prior history about CTh progression. The prediction accuracy of the proposed CTh prediction pipeline using a conditional score-based model was compared for sub-groups consisting of cognitively normal, mild cognitive impairment, and AD subjects. The Bland-Altman analysis shows our diffusion-based prediction model has a near-zero bias with narrow 95% confidential interval compared to the ground-truth CTh in 6-36 months. In addition, our conditional diffusion model has a stochastic generative nature, therefore, we demonstrated an uncertainty analysis of patient-specific CTh prediction through multiple realizations.

eess.IV↗

arXiv2023

Domain wall nature of sketched LaAlO3/SrTiO3 nanowires

The rich electron correlations and highly coherent transport in reconfigurable devices sketched by a conductive atomic force microscope tip at the LaAlO3/SrTiO3 interface have enabled the oxide platform an ideal playground for studying correlated electrons and quantum technological applications. Why these one-dimensional devices possess enhanced properties over the two-dimensional interface, however, has remained elusive. Here we provide evidence that one-dimensional LaAlO3/SrTiO3 nanowires are intrinsically ferroelastic domain walls by nature through thermodynamic study. We have observed spreading resistance anomalies under thermo-stimulus and temperature cycles, with characteristic temperatures matching domain wall polarity. This information is crucial in understanding the novel phenomena including superconductivity and high mobility quantum transport.

cond-mat.mes-hall↗

arXiv2023

Exploring Energy-based Language Models with Different Architectures and Training Methods for Speech Recognition

Energy-based language models (ELMs) parameterize an unnormalized distribution for natural sentences and are radically different from popular autoregressive language models (ALMs). As an important application, ELMs have been successfully used as a means for calculating sentence scores in speech recognition, but they all use less-modern CNN or LSTM networks. The recent progress in Transformer networks and large pretrained models such as BERT and GPT2 opens new possibility to further advancing ELMs. In this paper, we explore different architectures of energy functions and different training methods to investigate the capabilities of ELMs in rescoring for speech recognition, all using large pretrained models as backbones.

cs.CL↗

arXiv2022

Determination of building flood risk maps from LiDAR mobile mapping data

With increasing urbanization, flooding is a major challenge for many cities today. Based on forecast precipitation, topography, and pipe networks, flood simulations can provide early warnings for areas and buildings at risk of flooding. Basement windows, doors, and underground garage entrances are common places where floodwater can flow into a building. Some buildings have been prepared or designed considering the threat of flooding, but others have not. Therefore, knowing the heights of these facade openings helps to identify places that are more susceptible to water ingress. However, such data is not yet readily available in most cities. Traditional surveying of the desired targets may be used, but this is a very time-consuming and laborious process. This research presents a new process for the extraction of windows and doors from LiDAR mobile mapping data. Deep learning object detection models are trained to identify these objects. Usually, this requires to provide large amounts of manual annotations. In this paper, we mitigate this problem by leveraging a rule-based method. In a first step, the rule-based method is used to generate pseudo-labels. A semi-supervised learning strategy is then applied with three different levels of supervision. The results show that using only automatically generated pseudo-labels, the learning-based model outperforms the rule-based approach by 14.6% in terms of F1-score. After five hours of human supervision, it is possible to improve the model by another 6.2%. By comparing the detected facade openings' heights with the predicted water levels from a flood simulation model, a map can be produced which assigns per-building flood risk levels. This information can be combined with flood forecasting to provide a more targeted disaster prevention guide for the city's infrastructure and residential buildings.

cs.CV↗

arXiv2016

Calculating correlation coefficient for Gaussian copula

When Gaussian copula with linear correlation coefficient is used to model correlated random variables, one crucial issue is to determine a suitable correlation coefficient $ρ_z$ in normal space for two variables with correlation coefficient $ρ_x$. This paper attempts to address this problem. For two continuous variables, the marginal transformation is approximated by a weighted sum of Hermite polynomials, then, with Mehler's formula, a polynomial of $ρ_z$ is derived to approximate the function relationship between $ρ_x$ and $ρ_z$. If a discrete variable is involved, the marginal transformation is decomposed into piecewise continuous ones, and $ρ_x$ is expressed as a polynomial of $ρ_z$ by Taylor expansion. For a given $ρ_x$, $ρ_z$ can be efficiently determined by solving a polynomial equation.

stat.ME↗

arXiv2015

Generating correlated random vector by polynomial normal transformation

This paper develops a polynomial normal transformation model, whereby various non-normal probability distributions can be simulated by the standard normal distribution. Two methods are presented to determine the coefficients of polynomial model: (1) probability weighted moment (PWM) matching (2) percentile matching. Compared to the existing raw moment or L-moment matching, the proposed methods are more computationally convenient, and can be used to estimate the coefficients of polynomial model with a higher degree. Furthermore, for two correlated random variables, a polynomial equation is derived to estimate the equivalent correlation coefficient in standard normal space, and random vector with non-normal marginal distributions and prescribed correlation matrix can be generated. Finally, numerical examples are worked to demonstrate the proposed method.

stat.ME↗

arXiv2015

A method for calculating quantile function and its further use for data fitting

This paper introduces a polynomial transformation model based on Weibull distribution, whereby the analytical representation of the quantile function for many probability distributions can be obtained. Firstly, the target random variable $x$ with specified distribution is expressed as a polynomial of a Weibull random variable $z$, the coefficients are conveniently determined by the percentile matching method. Then, substituting $z$ with its quantile function $z=λ[-ln(1-u)]^{1/k}$ gives the analytical expression of the quantile function of $x$. Furthermore, using the probability weighted moments matching method, this polynomial transformation model can be used for data fitting. Through numerical experiment, it makes evident that the proposed model is capable of handling some distributions close to binomial which are difficult for the extant approaches, and the quantile functions of various distributions are accurately approximated within the probit range $[10^{-4},1-10^{-4}]$.

stat.ME↗