Xiaoguang Li
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Xiaoguang Li has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
- Tiejun Li
LMAM and School of Mathematical Sciences, Peking University, Beijing, China
5 shared publications - Mingguang Han
Laboratory of Mathematics and Applied Mathematics, School of Mathematical Sciences, Center for Machine Learning Research, Peking University, Beijing, P. R. China
2 shared publications - Yi Zeng
School of Mathematical Sciences, Center for Machine Learning Research, Laboratory of Mathematics and Applied Mathematics, Peking University, Beijing, P.R. China
2 shared publications - Chunhe Li
Institute of Science and Technology for Brain-Inspired Intelligence, Fudan University, Shanghai 200433, China
1 shared publication - Fugui Ma1 shared publication
- Tongkai Li
LMAM and School of Mathematical Sciences, Peking University 2 , Beijing, China
1 shared publication
Latest Publications
- Open Access
- 21 August 2026
- Published by Elsevier in SSRN Electronic Journal
iSMART: An Iterative Sampling-and-Regression Technique for Solving Martingale-Based PDEs
We propose the {\bf i}terative {\bf S}a{\bf M}pling-{\bf A}nd-{\bf R}egression {\bf T}echnique (iSMART) for high-dimensional martingale-based partial differential equations (PDEs) in this paper. By leveraging the $L^2$-projection property of conditional expectation and adopting the stop-gradient technique, iSMART reformulates the continuous martingale condition derived from PDEs into a sequence of tractable sampling-regression problems within an iterative framework. This approach relies solely on standard SDE path simulation and plain squared-error loss minimization, completely bypassing the need for adversarial optimization or nested expectation estimation in previous methods. iSMART accommodates linear, semi-linear, and fully nonlinear martingale-based PDEs within a unified iterative procedure. In particular, for fully nonlinear Hamilton-Jacobi-Bellman (HJB) equations, a freezing-and-compensating technique is introduced to strategically shift a portion of the nonlinearity into the SDE drift, thereby improving the convergence behavior of the iterations. Numerous numerical experiments on linear reaction-diffusion equations with sharp gradients, semilinear Burgers-type equations, and fully nonlinear HJB equations demonstrate the accuracy, efficiency, and robustness of the proposed approach in various high dimensions.
- 21 August 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessiSMART: An Iterative Sampling-and-Regression Technique for Solving Martingale-Based PDEs
We propose the {\bf i}terative {\bf S}a{\bf M}pling-{\bf A}nd-{\bf R}egression {\bf T}echnique (iSMART) for high-dimensional martingale-based partial differential equations (PDEs) in this paper. By leveraging the $L^2$-projection property of conditional expectation and adopting the stop-gradient technique, iSMART reformulates the continuous martingale condition derived from PDEs into a sequence of tractable sampling-regression problems within an iterative framework. This approach relies solely on standard SDE path simulation and plain squared-error loss minimization, completely bypassing the need for adversarial optimization or nested expectation estimation in previous methods. iSMART accommodates linear, semi-linear, and fully nonlinear martingale-based PDEs within a unified iterative procedure. In particular, for fully nonlinear Hamilton-Jacobi-Bellman (HJB) equations, a freezing-and-compensating technique is introduced to strategically shift a portion of the nonlinearity into the SDE drift, thereby improving the convergence behavior of the iterations. Numerous numerical experiments on linear reaction-diffusion equations with sharp gradients, semilinear Burgers-type equations, and fully nonlinear HJB equations demonstrate the accuracy, efficiency, and robustness of the proposed approach in various high dimensions.
- 22 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Signal Processing
- Vol. 74 (1053587X) , 2175-2191
Alternating Subspace Method for Sparse Recovery of Signals
Numerous renowned algorithms for tackling the compressed sensing problem employ an alternating strategy, which typically involves data matching in one module and denoising in another. We present a novel approach, the Alternating Subspace Method (ASM), which integrates the principles of the greedy methods (e.g., the orthogonal matching pursuit type methods) and the splitting methods (e.g., the approximate message passing type methods). Crucially, ASM enhances the splitting method by achieving fidelity in a subspace-restricted fashion. We reveal that such a restriction strategy guarantees global convergence via proximal residual control and establish its local geometric convergence on the LASSO problem. Numerical experiments on the LASSO, channel estimation, and dynamic compressed sensing problems demonstrate its high convergence rate and its capacity to incorporate different prior distributions. Overall, the proposed method is promising in terms of efficiency, accuracy, and flexibility, and has the potential to be competitive in different sparse recovery applications.
- 22 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Signal Processing
- Vol. 74 (1053587X) , 2175-2191
Alternating Subspace Method for Sparse Recovery of Signals
Numerous renowned algorithms for tackling the compressed sensing problem employ an alternating strategy, which typically involves data matching in one module and denoising in another. We present a novel approach, the Alternating Subspace Method (ASM), which integrates the principles of the greedy methods (e.g., the orthogonal matching pursuit type methods) and the splitting methods (e.g., the approximate message passing type methods). Crucially, ASM enhances the splitting method by achieving fidelity in a subspace-restricted fashion. We reveal that such a restriction strategy guarantees global convergence via proximal residual control and establish its local geometric convergence on the LASSO problem. Numerical experiments on the LASSO, channel estimation, and dynamic compressed sensing problems demonstrate its high convergence rate and its capacity to incorporate different prior distributions. Overall, the proposed method is promising in terms of efficiency, accuracy, and flexibility, and has the potential to be competitive in different sparse recovery applications.
- Open Access
- 16 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 212652-212671
QNOMP: A Quasi-Newton-Based Joint Multipath Optimization Method for Super-Resolution Channel Estimation and Extrapolation
Channel estimation and extrapolation are key challenges in massive multiple-input multiple-output communications. In this paper, we propose a quasi-Newton-based joint multipath optimization approach called Quasi-Newton Orthogonal Matching Pursuit (QNOMP) to address these problems. The algorithm comprises two stages for super-resolution recovery: first, we perform a computationally efficient on-grid OMP estimation of channel parameters in the sparsity domain (e.g., delay or angle); then, an off-grid optimization is carried out to achieve super-resolution. In the off-grid stage, we employ the quasi-Newton method to jointly estimate all multipath parameters, rather than optimizing them one at a time as in the Newtonized OMP approach. This joint estimation improves both computational efficiency and accuracy. Moreover, we derive an optimal extrapolation solution based on the linear minimum mean squared error criterion, establish its connection to the Slepian basis, and present a practical algorithm for extrapolation using the QNOMP results. We also propose a specialized approach that exploits the inherent block sparsity of the considered channels. Numerical experiments on simulated models and clustered delay line channels demonstrate the superior performance and low computational complexity of QNOMP.
- 16 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 212652-212671
Open AccessQNOMP: A Quasi-Newton-Based Joint Multipath Optimization Method for Super-Resolution Channel Estimation and Extrapolation
Channel estimation and extrapolation are key challenges in massive multiple-input multiple-output communications. In this paper, we propose a quasi-Newton-based joint multipath optimization approach called Quasi-Newton Orthogonal Matching Pursuit (QNOMP) to address these problems. The algorithm comprises two stages for super-resolution recovery: first, we perform a computationally efficient on-grid OMP estimation of channel parameters in the sparsity domain (e.g., delay or angle); then, an off-grid optimization is carried out to achieve super-resolution. In the off-grid stage, we employ the quasi-Newton method to jointly estimate all multipath parameters, rather than optimizing them one at a time as in the Newtonized OMP approach. This joint estimation improves both computational efficiency and accuracy. Moreover, we derive an optimal extrapolation solution based on the linear minimum mean squared error criterion, establish its connection to the Slepian basis, and present a practical algorithm for extrapolation using the QNOMP results. We also propose a specialized approach that exploits the inherent block sparsity of the considered channels. Numerical experiments on simulated models and clustered delay line channels demonstrate the superior performance and low computational complexity of QNOMP.
- 11 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Signal Processing
- Vol. 73 (1053587X) , 1759-1774
DMRA: An Adaptive Line Spectrum Estimation Method Through Dynamical Multiresolution of Atoms
We proposed a novel dense line spectrum super-resolution algorithm, the DMRA, that leverages dynamical multi-resolution of atoms technique to address the limitation of traditional compressed sensing methods when handling dense point-source signals. The algorithm utilizes a smooth tanh relaxation function to replace the l0 norm, promoting sparsity and jointly estimating the frequency atoms and complex gains. To reduce computational complexity and improve frequency estimation accuracy, a two-stage strategy was further introduced to dynamically adjust the number of the optimized degrees of freedom. Theoretical analysis were provided to validate the proposed method for multi-parameter estimations. DMRA presents excellent performance for the super-resolution of cluster-sparse signals, which is a typical scenario in different practical applications. It also outperforms the state-of-the-art methods in terms of frequency estimation accuracy and computational efficiency.
- 11 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Signal Processing
- Vol. 73 (1053587X) , 1759-1774
DMRA: An Adaptive Line Spectrum Estimation Method Through Dynamical Multiresolution of Atoms
We proposed a novel dense line spectrum super-resolution algorithm, the DMRA, that leverages dynamical multi-resolution of atoms technique to address the limitation of traditional compressed sensing methods when handling dense point-source signals. The algorithm utilizes a smooth tanh relaxation function to replace the l0 norm, promoting sparsity and jointly estimating the frequency atoms and complex gains. To reduce computational complexity and improve frequency estimation accuracy, a two-stage strategy was further introduced to dynamically adjust the number of the optimized degrees of freedom. Theoretical analysis were provided to validate the proposed method for multi-parameter estimations. DMRA presents excellent performance for the super-resolution of cluster-sparse signals, which is a typical scenario in different practical applications. It also outperforms the state-of-the-art methods in terms of frequency estimation accuracy and computational efficiency.
- 27 December 2021
- Published by AIP Publishing in The Journal of Chemical Physics
- Vol. 155 (24) , 245101
Uncovering the cell fate decision in lysis–lysogeny transition and stem cell development via Markov state modeling
Understanding the behavior of a complex gene regulatory network is a fundamental but challenging task in systems biology. How to reduce the large number of degrees of freedom of a specific network and identify its main biological pathway is the key issue. In this paper, we utilized the transition path theory (TPT) and Markov state modeling (MSM) framework to numerically study two typical cell fate decision processes: the lysis–lysogeny transition and stem cell development. The application of TPT to the lysis–lysogeny decision-making system reveals that the competitions of CI and Cro dimer binding play the major role in determining the cell fates. We also quantified the transition rates from the lysogeny to lysis state under different conditions. The overall computational results are consistent with biological intuitions but with more detailed information. For the stem cell developmental system, we applied the MSM to reduce the original dynamics to a moderate-size Markov chain. Further spectral analysis showed that the reduced system exhibits nine metastable states, which correspond to the refinement of the five known typical cell types in development. We further investigated the dominant transition pathways corresponding to the cell differentiation, reprogramming, and trans-differentiation. A similar approach can be applied to study other biological systems.
- 27 December 2021
- Published by AIP Publishing in The Journal of Chemical Physics
- Vol. 155 (24) , 245101
Uncovering the cell fate decision in lysis–lysogeny transition and stem cell development via Markov state modeling
Understanding the behavior of a complex gene regulatory network is a fundamental but challenging task in systems biology. How to reduce the large number of degrees of freedom of a specific network and identify its main biological pathway is the key issue. In this paper, we utilized the transition path theory (TPT) and Markov state modeling (MSM) framework to numerically study two typical cell fate decision processes: the lysis–lysogeny transition and stem cell development. The application of TPT to the lysis–lysogeny decision-making system reveals that the competitions of CI and Cro dimer binding play the major role in determining the cell fates. We also quantified the transition rates from the lysogeny to lysis state under different conditions. The overall computational results are consistent with biological intuitions but with more detailed information. For the stem cell developmental system, we applied the MSM to reduce the original dynamics to a moderate-size Markov chain. Further spectral analysis showed that the reduced system exhibits nine metastable states, which correspond to the refinement of the five known typical cell types in development. We further investigated the dominant transition pathways corresponding to the cell differentiation, reprogramming, and trans-differentiation. A similar approach can be applied to study other biological systems.