GZGuojun Zhang

Guojun Zhang

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Guojun Zhang has a total of 75 co-authors. Here are the top co-authors by number of shared publications:

  • Guanghui Zheng

    Clinical Laboratory of Beijing Tiantan Hospital, Capital Medical University, Beijing, People’s Republic of China

    9 shared publications
  • Hong Lv

    Clinical Laboratory of Beijing Tiantan Hospital, Capital Medical University, Beijing, People’s Republic of China

    7 shared publications
  • Lingye Qian

    Clinical Laboratory of Beijing Anzhen Hospital, Capital Medical University, Beijing, People’s Republic of China

    7 shared publications
  • Yijun Shi

    Clinical Laboratory of Beijing Tiantan Hospital, Capital Medical University, Beijing, People’s Republic of China

    7 shared publications
  • Defeng Sun

    Department of Applied Mathematics, The Hong Kong Polytechnic University, Hong Kong, China

    6 shared publications
  • Jialu Sun

    Clinical Laboratory of Beijing Tiantan Hospital, Capital Medical University, Beijing, People’s Republic of China

    5 shared publications

Latest Publications

  • Open Access

    dHPR: A Distributed Halpern Peaceman--Rachford Method for Non-smooth Distributed Optimization Problems

    by Zhangcheng Feng, Defeng Sun, Yancheng Yuan, Guojun Zhang

    This paper introduces the distributed Halpern Peaceman--Rachford (dHPR) method, an efficient algorithm for solving distributed convex composite optimization problems with non-smooth objectives, which achieves a non-ergodic $O(1/k)$ iteration complexity regarding Karush--Kuhn--Tucker residual. By leveraging the symmetric Gauss--Seidel decomposition, the dHPR effectively decouples the linear operators in the objective functions and consensus constraints while maintaining parallelizability and avoiding additional large proximal terms, leading to a decentralized implementation with provably fast convergence. The superior performance of dHPR is demonstrated through comprehensive numerical experiments on distributed LASSO, group LASSO, and $L_1$-regularized logistic regression problems.

  • Top Cited Papers
    Open Access

    HPR-LP: An implementation of an HPR method for solving linear programming

    by Kaihuang Chen, Defeng Sun, Yancheng Yuan, Guojun Zhang, Xinyuan Zhao

    In this paper, we introduce an HPR-LP solver, an implementation of a Halpern Peaceman–Rachford (HPR) method with semi-proximal terms for solving linear programming (LP). The HPR method enjoys the iteration complexity of $$O(1/k)$$ O ( 1 / k ) in terms of the Karush–Kuhn–Tucker residual and the objective error. Based on the complexity results, we design an adaptive strategy of restart and penalty parameter update to improve the efficiency and robustness of the HPR method. We conduct extensive numerical experiments on different LP benchmark datasets using an NVIDIA A100-SXM4-80GB GPU in different stopping tolerances. Our solver’s Julia version achieves a 2.39x to 5.70x speedup measured by SGM10 on benchmark datasets with presolve ( 2.03x to 4.06x without presolve) over the award-winning solver PDLP with the tolerance of $$10^{-8}$$ 10 - 8 . The Julia implementation of HPR-LP is available for downloading at https://github.com/PolyU-IOR/HPR-LP .


  • Open Access

    On the Relationships among GPU-Accelerated First-Order Methods for Solving Linear Programming

    by Kaihuang Chen, Defeng Sun, Yancheng Yuan, Guojun Zhang, Xinyuan Zhao

    This paper aims to understand the relationships among recently developed GPU-accelerated first-order methods (FOMs) for linear programming (LP), with particular emphasis on HPR-LP -- a Halpern Peaceman--Rachford (HPR) method for LP. Our findings can be summarized as follows: (i) the base algorithm of cuPDLPx, a recently released GPU solver, is a special case of the base algorithm of HPR-LP, thereby showing that cuPDLPx is another concrete implementation instance of HPR-LP; (ii) once the active sets have been identified, HPR-LP and EPR-LP -- an ergodic PR method for LP -- become equivalent under the same initialization; and (iii) extensive numerical experiments on benchmark datasets demonstrate that HPR-LP achieves the best overall performance among current GPU-accelerated LP solvers. These findings provide a strong motivation for using the HPR method as a baseline to further develop GPU-accelerated LP solvers and beyond.

  • Open Access

    HPR-QP: A dual Halpern Peaceman-Rachford method for solving large-scale convex composite quadratic programming

    by Kaihuang Chen, Defeng Sun, Yancheng Yuan, Guojun Zhang, Xinyuan Zhao

    In this paper, we introduce HPR-QP, a dual Halpern Peaceman-Rachford (HPR) method designed for solving large-scale convex composite quadratic programming. One distinctive feature of HPR-QP is that, instead of working with the primal formulations, it builds on the novel restricted Wolfe dual introduced in recent years. It also leverages the symmetric Gauss-Seidel technique to simplify subproblem updates without introducing auxiliary slack variables that typically lead to slow convergence. By restricting updates to the range space of the Hessian of the quadratic objective function, HPR-QP employs proximal operators of smaller spectral norms to speed up the convergence. Shadow sequences are elaborately constructed to deal with the range space constraints. Additionally, HPR-QP incorporates adaptive restart and penalty parameter update strategies, derived from the HPR method's $O(1/k)$ convergence in terms of the Karush-Kuhn-Tucker residual, to further enhance its performance and robustness. Extensive numerical experiments on benchmark data sets using a GPU demonstrate that our Julia implementation of HPR-QP significantly outperforms state-of-the-art solvers in both speed and scalability.

  • Open Access

    Peaceman-Rachford Splitting Method Converges Ergodically for Solving Convex Optimization Problems

    by Kaihuang Chen, Defeng Sun, Yancheng Yuan, Guojun Zhang, Xinyuan Zhao

    In this paper, we prove that the ergodic sequence generated by the Peaceman-Rachford (PR) splitting method with semi-proximal terms converges for convex optimization problems (COPs). Numerical experiments on the linear programming benchmark dataset further demonstrate that, with a restart strategy, the ergodic sequence of the PR splitting method with semi-proximal terms consistently outperforms both the point-wise and ergodic sequences of the Douglas-Rachford (DR) splitting method. These findings indicate that the restarted ergodic PR splitting method is a more effective choice for tackling large-scale COPs compared to its DR counterparts.

  • Open Access

    Clinical Characteristics and Predictors of Mortality of Patients with Post-Neurosurgical Meningitis-A 900-Cases Cohort Study

    by Guanghui Zheng, Yijun Shi, Jialu Sun, Siqi Wang, Lingye Qian, Hong Lv, Guojun Zhang, Kelin Chen

    Aim: To express the clinical characteristics of patients with post-neurosurgical meningitis (PNM) and launch a survival analysis to screen mortality predictors. Methods: A cohort analysis containing more than 70000 patients was evaluated, and all of them received neurosurgical procedure. Clinical and microbial epidemiology, therapy and mortality of PNM patients were reviewed. Multi-variable Cox proportional hazard models were applied to achieve survival analysis. Results: About 900 PNM patients from 3244 cases were selected for characteristics and survival analysis, the mean age of them was 41 (27– 54) years, 516 (57.3%) were men and 384 (42.7%) were women. The 28-day mortality was 12.4% (112 of 900) in patients with PNM. Hypertension, external ventricular drainage (EVD), and lumbar drainage (LD) are mortality predictors for PNM, with a hazard ratio (HR) of 2.641 (95% C.I. 1.563– 4.464, P< 0.001), 2.196 (95% C.I. 1.317– 3.662, P=0.003), and 1.818 (95% C.I. 1.126– 2.936, P=0.014). In treatment, the outcome of patients receiving three or more antibiotic combinations is better than that of patients receiving dual-drug combinations. Conclusion: The mortality of patients with PNM was relatively high, and the risk factors related to 28-days mortality were hypertension, EVD and LD and treatment with three or more antibiotics are much better.


  • Open Access

    Effect of antibiotic prophylaxis in the prognosis of Post-neurosurgical meningitis patients

    by Guanghui Zheng, Yijun Shi, Jialu Sun, Siqi Wang, Xiang Li, Hong Lv, Guojun Zhang

    Aims: To evaluate the effect of antibiotic prophylaxis(AP) in the prognosis of Post-neurosurgical meningitis(PNM) patients. Methods: A cohort analysis was performed using the clinical database in Beijing Tiantan Hospital and Capital Medical University. Data were collected on patients with the diagnosis of PNM (n = 3931) during 2012.01 to 2022.04. The microbial distribution, types of AP, and 42 and 90 days survival analysis of AP patients were evaluated using probable statistical methods. Independent risk factors for mortality were established by constructing a logistic regression analysis. Result: A total of 1,190 patients were included in this study, Klebsiella pneumoniae, Acinetobacter baumannii, and Staphylococcus aureus occupied the highest proportion. Of them, 929 cases received AP, cefuroxime and ceftriaxone are the most frequent used antibiotics. In addition, We found that PNM patients without AP significantly increased the 42 days and 90 days all-cause mortality rates. The use of different levels of AP did not improve patient outcomes, and ICU admission and assisted mechanical ventilation (AMV) were identified as independent mortality risk factors for PNM patient received AP. Conclusions: AP plays an important role in the prognosis of PNM patients and has a significant function in improving prognosis. The prevention of PNM with antibiotics prior to neurosurgery should be emphasized in clinical practice, and appropriate selection of antibiotics is necessary to prevent the occurrence of infection and inhibit the emergence of antibiotic-resistant bacteria.


  • Open Access

    An Efficient HPR Algorithm for the Wasserstein Barycenter Problem with $O({Dim(P)}/\varepsilon)$ Computational Complexity

    by Guojun Zhang, Yancheng Yuan, Defeng Sun

    In this paper, we propose and analyze an efficient Halpern-Peaceman-Rachford (HPR) algorithm for solving the Wasserstein barycenter problem (WBP) with fixed supports. While the Peaceman-Rachford (PR) splitting method itself may not be convergent for solving the WBP, the HPR algorithm can achieve an $O(1/\varepsilon)$ non-ergodic iteration complexity with respect to the Karush-Kuhn-Tucker (KKT) residual. More interestingly, we propose an efficient procedure with linear time computational complexity to solve the linear systems involved in the subproblems of the HPR algorithm. As a consequence, the HPR algorithm enjoys an $O({\rm Dim(P)}/\varepsilon)$ non-ergodic computational complexity in terms of flops for obtaining an $\varepsilon$-optimal solution measured by the KKT residual for the WBP, where ${\rm Dim(P)}$ is the dimension of the variable of the WBP. This is better than the best-known complexity bound for the WBP. Moreover, the extensive numerical results on both the synthetic and real data sets demonstrate the superior performance of the HPR algorithm for solving the large-scale WBP.

  • Open Access

    Nomogram Analysis of Clinical Characteristics and Mortality Risk Factor of Non-Fermentative Gram-Negative Bacteria-Induced Post-Neurosurgical Meningitis

    by Guanghui Zheng, Siqi Wang, Hong Lv, Guojun Zhang

    Objective: To explore the clinical characteristics of post-neurosurgical meningitis (PNM) patients infected with nonfermenting Gram-negative bacilli (NFGNB) and to evaluate the related mortality risk factors. Methods: A cohort analysis of PNM patients infected with NFGNB in Beijing Tiantan Hospital and Capital Medical University from 2012.1 to 2020.12. The microbial distribution, antimicrobial sensitivity and genotypes were tested, and potential mortality risk factors were evaluated using Mann–Whitney U or chi-squared tests. Independent risk factors for mortality were established by constructing a logistic model. Results: A total of 2940 PNM patients were enrolled in this study, of whom 207 (17.1%) were infected with NFGNB. Among these patients, 29 died of NFGNB meningitis, with an overall mortality rate of 14.0%. The top three NFGNBs were Acinetobacter baumannii (105 cases, 50.7%), Pseudomonas aeruginosa (29 cases, 14.0%) and Acinetobacter lwoffii (20 cases, 9.7%). Nomogram analysis revealed that hypertension (OR 4.551, 95% CI: 1.464– 14.154, P = 0.009), external ventricular drainage (EVD) (OR 3.944, 95% CI: 1.286– 12.095, P = 0.016), and assisted mechanical ventilator (AMV) (OR 6.192, 95% CI: 1.737– 22.081, P = 0.005) were independent risk factors for mortality. In addition, antibiotic prophylaxis was shown to play a vital role in NFGNB-induced PNM therapy. Conclusion: PNM patients infected with NFGNB have a high mortality rate. Hypertension, EVD and AMV were independent mortality risk factors, and clinical attention should be paid to their prevention and treatment.


  • Open Access

    Health care-associated meningitis: Clinical epidemiology and predictor for 90-day mortality

    by Guanghui Zheng, Yijun Shi, Yaowei Ding, Jialu Sun, Siqi Wang, Lingye Qian, Hong Lv, Guojun Zhang

    Aims To express the clinical predictor and epidemiology of patients with health care-associated meningitis (HAM) and launch a survival analysis to screen mortality predictors. Method A retrospective cohort analysis containing more than 70000 patients was evaluated. Clinical and microbial epidemiology, therapy and mortality of HAM patients were reviewed. Multivariable Cox proportional hazard models were applied to achieve survival analysis. Result 900 HAM patients from 3244 cases with cerebrospinal fluid(CSF) culture positivity were selected for epidemiology and survival analysis. The 90-day mortality was 12.4% (112 of 900) in people with HAM. Hypertension, external ventricular drainage (EVD), and lumbar drainage (LD) are mortality predictors for HAM, with a hazard ratio (HR) of 2.097 (95% C.I. 1.308–3.361, P = 0.002), 1.709 (95% C.I. 1.119–2.611, P = 0.013), and 1.614 (95% C.I. 1.083–2.405, P = 0.019). Reoperation was a factor favoring mortality, with an HR of 0.615 (95% C.I. 0.397–0.951, P = 0.029). In treatment, the outcome of patients receiving three or more antibiotic combinations is better than that of patients receiving dual-drug combinations. Conclusion The mortality of patients with HAM was relatively high, and the most important factors related to 90-days survival were hypertension, EVD and LD and treatment with three or more antibiotics are much better.