arXiv subjects

Hongyun Liu

Publications and source records attributed to Hongyun Liu.

4 recordsLinked to original sources

arXiv2026

Simulating Node Manipulations in Gaussian Graphical Models: The GGMNIRA Framework for Continuous and Ordinal Psychological Network Data

Scientific Abstract: In psychological network analysis, centrality indices are commonly used to evaluate the importance of nodes within a network. However, centrality only captures the static topological position of a node, and there is no sufficient theoretical justification for assuming that it reflects a node's influence on network dynamics. The NodeIdentifyR Algorithm (NIRA) offers an alternative by systematically applying simulated manipulations to node intercepts within the Ising model to evaluate nodes' projected importance, but this algorithm is restricted to binary data, and the manipulated parameter lacks a clear theoretical meaning outside the context of psychopathology. To address these limitations, we propose the Gaussian Graphical Model NodeIdentifyR Algorithm (GGMNIRA), which manipulates a node's conditional mean and uses Kullback-Leibler (KL) divergence to quantify the change in network distribution before and after manipulation, thereby extending this simulated manipulation logic to the Gaussian graphical model framework, which is applicable to continuous and ordinal data. Around this algorithm, we further developed a correlation stability coefficient and a nonparametric bootstrap difference test for KL divergence, with corresponding interpretive thresholds established through simulation studies. The framework was also extended to bridge Gaussian graphical models and moderated Gaussian graphical models, enabling its application to multi-construct comorbidity networks and to contexts involving moderation effects. All methods are implemented in the R package "GGMNIRA".

stat.ME↗

arXiv2026

Two-stage Estimation of Latent Variable Regression Models: A General, Root-N Consistent Solution

Latent variable (LV) models are widely used in psychological research to investigate relationships among unobservable constructs. When one-stage estimation of the overall LV model is challenging, two-stage factor score regression (FSR) serves as a convenient alternative: the measurement model is fitted to obtain factor scores in the first stage, which are then used to fit the structural model in the subsequent stage. However, naive application of FSR is known to yield biased estimates of structural parameters. In this paper, we develop a generic bias-correction framework for two-stage estimation of parametric statistical models and tailor it specifically to FSR. Unlike existing bias-corrected FSR solutions, the proposed method applies to a broader class of LV models and does not require computing specific types of factor scores. We establish the root-n consistency of the proposed bias-corrected two-stage estimator under mild regularity conditions. To ensure broad applicability and minimize reliance on complex analytical derivations, we introduce a stochastic approximation algorithm for point estimation and a Monte Carlo-based procedure for variance estimation. In a sequence of Monte Carlo experiments, we demonstrate that the bias-corrected FSR estimator performs comparably to the ``gold standard'' one-stage maximum likelihood estimator. These results suggest that our approach offers a straightforward yet effective alternative for estimating LV models.

stat.ME↗

arXiv2025

A Topological Improvement of the Overall Performance of Sparse Evolutionary Training: Motif-Based Structural Optimization of Sparse MLPs Project

Deep Neural Networks (DNNs) have been proven to be exceptionally effective and have been applied across diverse domains within deep learning. However, as DNN models increase in complexity, the demand for reduced computational costs and memory overheads has become increasingly urgent. Sparsity has emerged as a leading approach in this area. The robustness of sparse Multi-layer Perceptrons (MLPs) for supervised feature selection, along with the application of Sparse Evolutionary Training (SET), illustrates the feasibility of reducing computational costs without compromising accuracy. Moreover, it is believed that the SET algorithm can still be improved through a structural optimization method called motif-based optimization, with potential efficiency gains exceeding 40% and a performance decline of under 4%. This research investigates whether the structural optimization of Sparse Evolutionary Training applied to Multi-layer Perceptrons (SET-MLP) can enhance performance and to what extent this improvement can be achieved.

cs.NE↗

arXiv2022

FIRED: a fine-grained robust performance diagnosis framework for cloud applications

To run a cloud application with the required service quality, operators have to continuously monitor the cloud application's run-time status, detect potential performance anomalies, and diagnose the root causes of anomalies. However, existing models of performance anomaly detection often suffer from low re-usability and robustness due to the diversity of system-level metrics being monitored and the lack of high-quality labeled monitoring data for anomalies. Moreover, the current coarse-grained analysis models make it difficult to locate system-level root causes of the application performance anomalies for effective adaptation decisions. We provide a FIne-grained Robust pErformance Diagnosis (FIRED) framework to tackle those challenges. The framework offers an ensemble of several well-selected base models for anomaly detection using a deep neural network, which adopts weakly-supervised learning considering fewer labels exist in reality. The framework also employs a real-time fine-grained analysis model to locate dependent system metrics of the anomaly. Our experiments show that the framework can achieve the best detection accuracy and algorithm robustness, and it can predict anomalies in four minutes with F1 score higher than 0.8. In addition, the framework can accurately localize the first root causes, and with an average accuracy higher than 0.7 of locating first four root causes.

cs.DC↗