Author name cluster

Qing Zhou

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

16 papers

2 author rows

AAAI Conference 2026 Conference Paper

Reasoning via Implicit Self-supervised Emergence for Instruction Segmentation

Qing Zhou
Lichang Yang
Yuyu Jia
Junyu Gao
Weiping Ni
Junzheng Wu
Qi Wang

We challenge the assumption that complex instruction-guided segmentation tasks necessitate equally complex and explicit supervision. This paper introduces RISE (Reasoning via Implicit Self-supervised Emergence), a framework that learns intricate compositional reasoning, spanning spatial relations to world knowledge, without a single ground-truth mask. To achieve this, RISE employs reinforcement learning with GRPO guided by a single, strikingly simple reward: the semantic alignment score between the textual instruction and the predicted image region. Our primary discovery is the implicit emergence of a high-quality chain-of-thought process from this minimalist signal. Within a structured format, the model autonomously learns to understand instructions by accessing its latent knowledge, inferring spatial relationships—capabilities inherent in its architecture but unlocked by our simple objective. Remarkably, our emergent reasoning yields highly competitive results: RISE achieves 58.7 gIoU on the ReasonSeg benchmark, on par with methods using geometric rewards. Furthermore, we show extreme data efficiency: a variant trained on only 2,000 ImageNet-label pairs establishes a new state-of-the-art for annotation-free referring segmentation with 79.6 cIoU on RefCOCO.

PDF Details DOI

EAAI Journal 2026 Journal Article

Superpixel-based scribble-level supervision diffusion for the interactive segmentation of multi-configuration chips

Hang Zhang
Qing Zhou
Weidong Tang
Shengfeng Chen
Yuanqiang Luo

The precise segmentation of functional regions is imperative for the intelligent manufacturing of light emitting diode (LED) chips. However, the performance of existing precise segmentation methods is often contingent on the full density supervision, a process that is profoundly labour-intensive. The multi-configuration of LED chips greatly increases the workload of supervision acquisition. To overcome this challenge, a superpixel-based scribble-level supervision diffusion (SSSD) is proposed for the interactive segmentation of LED chips. The SSSD integrates the excellent boundary identification capability of superpixel method with the general semantic extraction ability of foundation model, which is capable of achieving precise segmentation through the probabilistic diffusion of scribble-level supervision. Specifically, a hierarchical simple linear iterative clustering and merging is utilised to generate superpixels, which are capable of identifying the boundaries of functional regions. Furthermore, in order to extract the general and robust features for LED chips, a Learning Robust Visual Features without Supervision model (DINOv2) is employed. Finally, a scribble-level supervision diffusion is developed to achieve the precise segmentation. The results demonstrate the efficacy of SSSD in precisely segmenting the functional regions of LED chips. The Dice values of SSSD across LED chips with three configurations are 97. 04%, 95. 95% and 96. 86%, respectively.

Details DOI

NeurIPS Conference 2025 Conference Paper

Proper Hölder-Kullback Dirichlet Diffusion: A Framework for High Dimensional Generative Modeling

WanPeng Zhang
Yuhao Fang
Xihang Qiu
Jiarong Cheng
Jialong Hong
Bin Zhai
Qing Zhou
Yao Lu

Diffusion-based generative models have long depended on Gaussian priors, with little exploration of alternative distributions. We introduce a Proper Hölder-Kullback Dirichlet framework that uses time-varying multiplicative transformations to define both forward and reverse diffusion processes. Moving beyond conventional reweighted evidence lower bounds (ELBO) or Kullback–Leibler upper bounds (KLUB), we propose two novel divergence measures: the Proper Hölder Divergence (PHD) and the Proper Hölder–Kullback (PHK) divergence, the latter designed to restore symmetry missing in existing formulations. When optimizing our Dirichlet diffusion model with PHK, we achieve a Fréchet Inception Distance (FID) of 2. 78 on unconditional CIFAR-10. Comprehensive experiments on natural-image datasets validate the generative strengths of model and confirm PHK’s effectiveness in model training. These contributions expand the diffusion-model family with principled non-Gaussian processes and effective optimization tools, offering new avenues for versatile, high-fidelity generative modeling.

PDF Details

JMLR Journal 2024 Journal Article

Learning Gaussian DAGs from Network Data

Hangjian Li
OSCAR HERNAN MADRID PADILLA
Qing Zhou

Structural learning of directed acyclic graphs (DAGs) or Bayesian networks has been studied extensively under the assumption that the data are independent. We propose a new Gaussian DAG model for dependent data which assumes the observations are correlated according to an undirected network. Under this model, we develop a method to estimate the DAG structure given a topological ordering of the nodes. The proposed method jointly estimates the Bayesian network and the correlations among observations by optimizing a scoring function based on penalized likelihood. We show that under some mild conditions, the proposed method produces consistent estimators after one iteration. Extensive numerical experiments also demonstrate that, by jointly estimating the DAG structure and the sample correlation, our method achieves much higher accuracy in structure learning. When the node ordering is unknown, through experiments on synthetic and real data, we show that our algorithm can be used to estimate the correlations between samples, with which we can de-correlate the dependent data to significantly improve the performance of classical DAG learning methods. [abs] [ pdf ][ bib ] &copy JMLR 2024. ( edit, beta )

PDF Details

TMLR Journal 2023 Journal Article

Bayesian Causal Bandits with Backdoor Adjustment Prior

Jireh Huang
Qing Zhou

The causal bandit problem setting is a sequential decision-making framework where actions of interest correspond to interventions on variables in a system assumed to be governed by a causal model. The underlying causality may be exploited when investigating actions in the interest of optimizing the yield of the reward variable. Most existing approaches assume prior knowledge of the underlying causal graph, which is in practice restrictive and often unrealistic. In this paper, we develop a novel Bayesian framework for tackling causal bandit problems that does not rely on possession of the causal graph, but rather simultaneously learns the causal graph while exploiting causal inferences to optimize the reward. Our methods efficiently utilize joint inferences from interventional and observational data in a unified Bayesian model constructed with intervention calculus and causal graph learning. For the implementation of our proposed methodology in the discrete distributional setting, we derive an approximation of the sampling variance of the backdoor adjustment estimator. In the Gaussian setting, we characterize the interventional variance with intervention calculus and propose a simple graphical criterion to share information between arms. We validate our proposed methodology in an extensive empirical study, demonstrating compelling cumulative regret performance against state-of-the-art standard algorithms as well as optimistic implementations of their causal variants that assume strong prior knowledge of the causal structure.

PDF Details

TIST Journal 2023 Journal Article

Cost-sensitive Tensor-based Dual-stage Attention LSTM with Feature Selection for Data Center Server Power Forecasting

Ziyu Shen
Binghui Liu
Qing Zhou
Zheng Liu
Bin Xia
Yun Li

Power forecasting has a guiding effect on power-aware scheduling strategies to reduce unnecessary power consumption in data centers. Many metrics related to power consumption can be collected in physical servers, such as the status of CPU, memory, and other components. However, most existing methods empirically exploit a small number of metrics to forecast power consumption. To this end, this article uses feature selection based on causality to explore the metrics that strongly influence the power consumption of different tasks. Moreover, we propose a tensor-based dual-stage attention LSTM to forecast the non-linear and non-periodic power consumption. In the proposed model, a multi-way delay embedding transform is utilized to convert the time series into tensors along the temporal direction. The LSTM combines with the tensor technique and the attention mechanism to capture the temporal pattern effectively. In addition, we adopt the cost-sensitive loss function to optimize the specific power forecasting problem in data centers. The experimental results demonstrate that our method can achieve up to 1.4% to 4.3% forecasting accuracy improvement compared with the state-of-the-art models.

Details DOI

YNICL Journal 2023 Journal Article

MRI histogram analysis of tumor-infiltrating CD8+ T cell levels in patients with glioblastoma

Caiqiang Xue
Qing Zhou
Peng Zhang
Bin Zhang
Qiu Sun
Shenglin Li
Juan Deng
Xianwang Liu

OBJECTIVE: To investigate the utility of preoperative magnetic resonance imaging histogram analysis for evaluating tumor-infiltrating CD8+ T cells in patients with glioblastoma (GBM). METHODS: We retrospectively analyzed the pathological and imaging data of 61 patients with GBM confirmed by surgery and pathology. Moreover, the levels of tumor-infiltrating CD8+ T cells in tumor tissue samples obtained from the patients were quantified through immunohistochemical staining and evaluated with respect to overall survival. The patients were divided into the high and low CD8 expression groups. Preoperative T1-weighted contrast-enhanced (T1C) histogram parameters of patients with GBM were extracted using Firevoxel software. We investigated the correlation between the histogram feature parameters and CD8+ T cells. We performed statistical analyses of the T1C histogram parameters in both groups and identified characteristic parameters with significant between-group differences. Additionally, we performed a receiver operating characteristic curve (ROC) analysis to determine the predictive utility of these parameters. RESULTS: The levels of tumor-infiltrating CD8+ T cells were positively associated with overall survival in patients with GBM (P = 0.0156). Among the T1C histogram features, the mean, 5th, 10th, 25th, and 50th percentiles were negatively correlated with the levels of CD8+ T cells. Moreover, the coefficient of variation (CV) was positively correlated with the levels of CD8+ T cells (all P < 0.05). There was a significant between-group difference in the CV, 1st, 5th, 10th, 25th, and 50th percentiles (all p < 0.05). The ROC curve analysis revealed that the CV had the highest AUC value (0.783; 95% confidence interval: 0.658-0.878), with sensitivity and specificity values of 0.784 and 0.750, respectively, for distinguishing between the groups. CONCLUSIONS: The preoperative T1C histogram have additional value for the levels of tumor-infiltrating CD8+ T cells in patients with GBM.

Details DOI

ICML Conference 2023 Conference Paper

Structure Learning of Latent Factors via Clique Search on Correlation Thresholded Graphs

Dale Kim
Qing Zhou

Despite the widespread application of latent factor analysis, existing methods suffer from the following weaknesses: requiring the number of factors to be known, lack of theoretical guarantees for learning the model structure, and nonidentifiability of the parameters due to rotation invariance properties of the likelihood. We address these concerns by proposing a fast correlation thresholding (CT) algorithm that simultaneously learns the number of latent factors and a rotationally identifiable model structure. Our novel approach translates this structure learning problem into the search for so-called independent maximal cliques in a thresholded correlation graph that can be easily constructed from the observed data. Our clique analysis technique scales well up to thousands of variables, while competing methods are not applicable in a reasonable amount of running time. We establish a finite-sample error bound and high-dimensional consistency for the structure learning of our method. Through a series of simulation studies and a real data example, we show that the CT algorithm is an accurate method for learning the structure of factor analysis models and is robust to violations of its assumptions.

Details

YNICL Journal 2022 Journal Article

Radiomics for prediction of intracerebral hemorrhage outcomes: A retrospective multicenter study

Xiaoyu Huang
Dan Wang
Qiaoying Zhang
Yaqiong Ma
Hui Zhao
Shenglin Li
Juan Deng
Jialiang Ren

BACKGROUND: Accurate risk stratification of patients with intracerebral hemorrhage (ICH) could help refine adjuvant therapy selection and better understand the clinical course. We aimed to evaluate the value of radiomics features from hematomal and perihematomal edema areas for prognosis prediction and to develop a model combining clinical and radiomic features for accurate outcome prediction of patients with ICH. METHODS: This multicenter study enrolled patients with ICH from January 2016 to November 2021. Their outcomes at 3 months were recorded based on the modified Rankin Scale (good, 0-3; poor, 4-6). Independent clinical and radiomic risk factors for poor outcome were identified through multivariate logistic regression analysis, and predictive models were developed. Model performance and clinical utility were evaluated in both internal and external cohorts. RESULTS: Among the 1098 ICH patients evaluated (mean age, 60 ± 13 years), 703 (64 %) had poor outcomes. Age, hemorrhage volume and location, and Glasgow Coma Scale (GCS) were independently associated with outcomes. The area under the receiver operating characteristic curve (AUC) of the clinical model was 0.881 in the external validation cohort. Addition of the Rad-score (combined hematoma and perihematomal edema area) improved predictive accuracy and model performance (AUC, 0.893), net reclassification improvement, 0.140 (P < 0.001), and integrated discrimination improvement, 0.050 (P < 0.001). CONCLUSIONS: The radiomics features of hematomal and perihematomal edema area have additional value in prognostic prediction; moreover, addition of radiomic features significantly improves model accuracy.

Details DOI

TMLR Journal 2022 Journal Article

Sequentially learning the topological ordering of directed acyclic graphs with likelihood ratio scores

Gabriel Ruiz
OSCAR HERNAN MADRID PADILLA
Qing Zhou

Causal discovery, the learning of causality in a data mining scenario, has been of strong scientific and theoretical interest as a starting point to identify "what causes what?'' Contingent on assumptions and a proper learning algorithm, it is sometimes possible to identify and accurately estimate an underlying directed acyclic graph (DAG), as opposed to a Markov equivalence class of graphs that gives ambiguity of causal directions. The focus of this paper is in highlighting the identifiability and estimation of DAGs through a sequential sorting procedure that orders variables one at a time, starting at root nodes, followed by children of the root nodes, and so on until completion. We demonstrate a novel application of this general sequential approach to estimate the topological ordering of the DAG corresponding to a linear structural equation model with a non-Gaussian error distribution family. At each step of the procedure, only simple likelihood ratio scores are calculated on regression residuals to decide the next node to append to the current partial ordering. The computational complexity of our algorithm on a $p$-node problem is $\mathcal{O}(pd)$, where $d$ is the maximum neighborhood size. Under mild assumptions, the population version of our procedure provably identifies a true ordering of the underlying DAG. We provide extensive numerical evidence to demonstrate that this sequential procedure scales to possibly thousands of nodes and works well for high-dimensional data. We accompany these numerical experiments with an application to a single-cell gene expression dataset. Our $\texttt{R}$ package with examples and installation instructions can be found at https://gabriel-ruiz.github.io/scorelingam/.

PDF Details

JMLR Journal 2020 Journal Article

Learning Big Gaussian Bayesian Networks: Partition, Estimation and Fusion

Jiaying Gu
Qing Zhou

Structure learning of Bayesian networks has always been a challenging problem. Nowadays, massive-size networks with thousands or more of nodes but fewer samples frequently appear in many areas. We develop a divide-and-conquer framework, called partition-estimation-fusion (PEF), for structure learning of such big networks. The proposed method first partitions nodes into clusters, then learns a subgraph on each cluster of nodes, and finally fuses all learned subgraphs into one Bayesian network. The PEF method is designed in a flexible way so that any structure learning method may be used in the second step to learn a subgraph structure as either a DAG or a CPDAG. In the clustering step, we adapt hierarchical clustering to automatically choose a proper number of clusters. In the fusion step, we propose a novel hybrid method that sequentially adds edges between subgraphs. Extensive numerical experiments demonstrate the competitive performance of our PEF method, in terms of both speed and accuracy compared to existing methods. Our method can improve the accuracy of structure learning by 20% or more, while reducing running time up to two orders-of-magnitude. [abs] [ pdf ][ bib ] &copy JMLR 2020. ( edit, beta )

PDF Details

NeurIPS Conference 2019 Conference Paper

Globally optimal score-based learning of directed acyclic graphs in high-dimensions

Bryon Aragam
Arash Amini
Qing Zhou

We prove that $\Omega(s\log p)$ samples suffice to learn a sparse Gaussian directed acyclic graph (DAG) from data, where $s$ is the maximum Markov blanket size. This improves upon recent results that require $\Omega(s^{4}\log p)$ samples in the equal variance case. To prove this, we analyze a popular score-based estimator that has been the subject of extensive empirical inquiry in recent years and is known to achieve state-of-the-art results. Furthermore, the approach we study does not require strong assumptions such as faithfulness that existing theory for score-based learning crucially relies on. The resulting estimator is based around a difficult nonconvex optimization problem, and its analysis may be of independent interest given recent interest in nonconvex optimization in machine learning. Our analysis overcomes the drawbacks of existing theoretical analyses, which either fail to guarantee structure consistency in high-dimensions (i. e. learning the correct graph with high probability), or rely on restrictive assumptions. In contrast, we give explicit finite-sample bounds that are valid in the important $p\gg n$ regime.

PDF Details

JMLR Journal 2015 Journal Article

Concave Penalized Estimation of Sparse Gaussian Bayesian Networks

Bryon Aragam
Qing Zhou

We develop a penalized likelihood estimation framework to learn the structure of Gaussian Bayesian networks from observational data. In contrast to recent methods which accelerate the learning problem by restricting the search space, our main contribution is a fast algorithm for score-based structure learning which does not restrict the search space in any way and works on high-dimensional data sets with thousands of variables. Our use of concave regularization, as opposed to the more popular $\ell_0$ (e.g. BIC) penalty, is new. Moreover, we provide theoretical guarantees which generalize existing asymptotic results when the underlying distribution is Gaussian. Most notably, our framework does not require the existence of a so-called faithful DAG representation, and as a result, the theory must handle the inherent nonidentifiability of the estimation problem in a novel way. Finally, as a matter of independent interest, we provide a comprehensive comparison of our approach to several standard structure learning methods using open-source packages developed for the R language. Based on these experiments, we show that our algorithm obtains higher sensitivity with comparable false discovery rates for high- dimensional data and scales efficiently as the number of nodes increases. In particular, the total runtime for our method to generate a solution path of 20 estimates for DAGs with 8000 nodes is around one hour. [abs] [ pdf ][ bib ] &copy JMLR 2015. ( edit, beta )

PDF Details

EAAI Journal 2011 Journal Article

Modeling of the carbon dioxide capture process system using machine intelligence approaches

Qing Zhou
Yuxiang Wu
Christine W. Chan
Paitoon Tontiwachwuthikul

Improving the efficiency of the carbon dioxide (CO2) capture process requires a good understanding of the intricate relationships among parameters involved in the process. The objective of this paper is to study the relationships among the significant parameters impacting CO2 production. An enhanced understanding of the intricate relationships among the process parameters supports prediction and optimization, thereby improving efficiency of the CO2 capture process. Our modeling study used the 3-year operational data collected from the amine-based post combustion CO2 capture process system at the International Test Centre (ITC) of CO2 Capture located in Regina, Saskatchewan of Canada. This paper describes the data modeling process using the approaches of (1) neural network modeling combined with sensitivity analysis and (2) neuro-fuzzy modeling technique. The results from the two modeling processes were compared from the perspectives of predictive accuracy, inclusion of parameters, and support for explication of problem space. We conclude from the study that the neuro-fuzzy modeling technique was able to achieve higher accuracy in predicting the CO2 production rate than the combined approach of neural network modeling and sensitivity analysis.

Details DOI

EAAI Journal 2010 Journal Article

A comparison of two data analysis techniques and their applications for modeling the carbon dioxide capture process

Yuxiang Wu
Qing Zhou
Christine W. Chan

Improving the efficiency of the carbon dioxide (CO2) capture process requires a good understanding of the intricate relationships among parameters involved in the process. The objective of this research is to study the nature of relationships among the key parameters using the approaches of artificial neural network and statistical analysis. Our modeling study used the three-year operational data collected from the amine-based post-combustion CO2 capture process at the International Test Centre of CO2 Capture (ITC) located in Regina, Saskatchewan of Canada. The goal of CO2 capture is to capture and remove CO2 from industrial gas streams before they are released into the atmosphere. The amine solution is used at ITC for absorbing CO2 from the industrial flue gas, and then the CO2 is separated from the amine solution. The amine solution recycles for further CO2 capture and the CO2 stream can be stored or used for other industrial purposes. This paper describes the data modeling process using the approaches of: (1) statistical analysis and (2) neural network modeling combined with sensitivity analysis. The results from the two modeling process were compared from the perspectives of predictive accuracy, inclusion of parameters, support for exploration and explication of problem space, modeling uncertainty, and involvement of experts. It was observed that the approach of neural network modeling combined with sensitivity analysis achieved much higher accuracy on predicting CO2 production rate than the statistical study.

Details DOI

TCS Journal 2000 Journal Article

Grzegorczyk's hierarchy of computable analysis

Qing Zhou

This paper deals with the computability in analysis within the framework of Grzegorczyk's hierarchy, which is in the number 1 of addendum of open problems in Pour-El and Richards ([5], Computability in Analysis and Physics, Springer, Berlin, 1989). We combine two concepts, computability for sequences of real-valued functions and Grzegorczyk's hierarchy for recursive number theoretic functions, together and examine the computability in analysis restricted to primitive recursion and below. The notions of ( E r ) primitive computability structures on Banach space, in particular, for sequences of reals and real-valued functions are introduced; relations between ( E r ) primitive computability structures are proved; some basic properties are studied.

Details DOI