第一场
报告承办单位: 数学与统计学院
报 告 题 目 : Robust Log-Contrast Regression for High-Dimensional Compositional Microbiome Data with Outliers
报告人姓名: 肖运海 教授
报告人所在单位: 河南大学
报告人职称: 教授、博士生导师
报告时间: 2026 年 11 月 8 日 星期日 上午 9:30
报告地点: 理科楼 A212
报告摘要: It is well known that microbiome sequencing data are naturally compositional and often have a complex subcomposition structure, as well as a lot of noise and some outliers. These characteristics make classical log-contrast regression difficult to apply, because it is sensitive to extreme values and cannot fully meet the linear constraints of compositional predictors.To tackle these issues, in this paper, we propose a robust log-contrast regression framework that incorporates the Huber loss func ion along with a Lasso penalty, which not only boosts resistance to outliers but also facilitates sparse variable selection.It is important to note that the general linear constraints, which are explicitly imposed, serve to preserve the algebraic coherence of compositional and subcompositional covariates.Under some mild assumptions, we theoretically prove that the proposed method achieves non-asymptotic guarantees in finite samples, comparable to those of Huber loss estimators, ensuring robustness.To implement the method efficiently, we develop a symmetric Gauss--Seidel-based alternating direction method of multipliers based on its dual formulation with a convergence guarantee, and it is also easy to implement in the sense that each subproblem admits a closed-form solution.To evaluate the statistical performance of the proposed method, we conduct a series of numerical experiments using simulated data and gut microbiome data, which show that the proposed method achieves better estimation accuracy, competitive prediction performance, and more reliable outlier identification than the benchmark method RobRegCC.
报告人简介:肖运海,河南大学教授,博士生导师,河南省特聘教授,河南省杰出青年基金获得者,研究方向为最优化算法及应用。2001 年本科毕业于河南师范大学,2004 年获广西大学硕士学位,2007 年获湖南大学博士学位,并分别于 2010 年和 2011 年在南京大学和台湾成功大学完成博士后研究。曾在加拿大西蒙弗雷泽大学、新加坡国立大学、香港理工大学和台湾成功大学等进行学术交流访问。主持国家自然科学基金项目 4 项(面上 3 项,青年 1 项),中国博士后面上项目 1 项,参加 973 计划项目 1 项。在MPC、JCGS、COAP、JSC、CSDA 等学术期刊上发表论文60 余篇,被引约 1700 次。担任中国运筹学会理事和宣传工作委员会副主任,中国工业与应用数学学会理事和青年工作委员会副主任,中国运筹学会数学规划分会常务理事、算法软件与应用分会常务理事,河南省运筹学会副理事长,河南省统筹法优选法与经济数学研究会副理事长,河南大学第五届学术委员会委员,曾任中国运筹学会副秘书长等。
第二场
报告承办单位: 数学与统计学院
报告题目: Inertial Viscosity Algorithms for Variational Inclusion Problems
报告人姓名: 秦小龙 教授
报告人所在单位: 杭州师范大学
报告人职称: 教授、博士生导师
报告时间: 2026 年 11 月 8 日 星期日 上午 11:00
报告地点: 理科楼 A212
报告摘要:Variational inclusion is a fundamental mathematical model for numerous problems arising in real-world applications. It is closely connected to as split feasibility problems, variational inequalities, equilibrium problems and so on and has been extensively studied. In this talk, we present some inertial viscosity algorithms based on the forward-backward splitting technique for the solution of monotone inclusion problems in real Hilbert spaces. Our algorithms work adaptively in the absence of the Lipschitz constant of the single-valued operator involved. The weak and strong convergence of the algorithms are investigated and established under some appropriate assumptions. We also presented some numerical experiments to support our algorithms.
报告人简介:秦小龙,教授,博士生导师,省特聘专家,斯坦福前 2%学者,美国 Sigma Xi 正会员(Full Member),英国 IMA 会士(Fellow),自首届起连续九年入选科睿唯安(汤森路透)全球高被引,连续十一年入选爱思唯尔中国高被引,主要从事非线性分析,最优化理论与应用的研究。近年来已在 SIOPT, SICON, ESAIM-cocv, CVPDE, IEEE Trans等国际学术刊物发表论文一百余篇,获省自然科学奖三等奖一项,科睿唯安同行审阅奖两项(数学和跨学科),最佳论文奖两项(JOGO 和数学物理学报),担任 AMSA, JNVA, JOTA, JCAM, Opimization 等多个本领域国际期刊编委,并与 2025 受邀参加华人数学家大会(ICCM)作 45 分钟邀请报告。
第三场
报告承办单位: 数学与统计学院
报告题目: Conditional and Efficient Prediction Intervals via Shape Adaptive Conformal Prediction
报告人姓名: 孔令臣 教授
报告人所在单位: 北京交通大学
报告人职称: 教授、博士生导师
报告时间: 2026 年 11 月 8 日 星期日 下午 15:00-18:00
报告地点: 理科楼 A212
报告摘要: The standard distribution-free guarantee of conformal prediction is marginal and does not control coverage conditional on the covariates by itself. Moreover, coverage alone is insufficient: overly wide intervals may be valid but uninformative, so efficiency requires short intervals among those attaining a common coverage level. For a fixed pre-fitted regression function, the conditional signed-residual law governs both objectives: it determines the conditional coverage of any residual interval and the minimum length attainable by a connected interval at the prescribed coverage level. This observation motivates Shape-Adaptive Conformal Prediction (ShapeCP). ShapeCP represents each covariate value by fitted evaluations of its conditional residual CDF on a common threshold grid, localizes observations by profile similarity rather than covariate distance, and selects a shortest connected window from the resulting target-specific weighted residual distribution. Therefore it adapts interval location, width, and tail allocation before conformally calibrating endpoint violations on held-out data. This calibration guarantees finite-sample marginal coverage under exchangeability, regardless of profile-estimation accuracy. We establish a profile-to-law transport bound and convergence rates for the localized residual CDF and conformal correction under regularity conditions. These results yield asymptotic pointwise conditional coverage and convergence of interval length to the shortest connected conditional-oracle benchmark. Simulations and applications to GPU runtime and online news popularity demonstrate that ShapeCP combines near-nominal marginal coverage with competitive local coverage and short prediction intervals.
报告人简介:孔令臣,教授,博士生导师,中国运筹学会数学规划分会理事长,北京交通大学数学与统计学院副院长。主要从事对称锥互补问题和最优化、高维数据分析、统计优化与学习、医学成像等方面的研究。在《Mathematical Programming》《SIAM Journalon Optimization》《IEEE Transactions on Pattern Analysis and Machine Intelligence》《IEEE Transactions on Signal Processing》《Technometrics》《Statistica Sinica》《Electronic Journal of Statistics》等期刊发表论文 100 余篇,主持国家重点研发计划、国家自然科学基金重点项目及面上项目等 10 余项。获教育部自然科学二等奖、中国运筹学会青年奖及北京市高等教育教学成果一等奖等荣誉。担任《Asia-Pacific Journal of Operational Research》等期刊编委。
第四场
报告承办单位: 数学与统计学院
报告题目: 相关系数与预测能力
报告人姓名: 朱利平 教授
报告人所在单位: 中国人民大学
报告人职称: 教授、博士生导师
报告时间: 2026 年 11 月 8 日 星期日 下午 15:00-18:00
报告地点: 理科楼 A212
报告摘要: 相关系数是度量变量间依赖关系的基本工具。本报告首先从经典的皮尔逊相关系数出发,阐明其与线性模型预测能力之间的内在联系,即皮尔逊相关系数的平方恰好等于线性回归模型的可决系数。在此基础上,我们引入近年来由苏拉夫·查特吉提出的一种基于秩的新型相关性度量——查特吉相关系数。该系数能够捕捉变量间任意形式的依赖关系,且在总体层面满足:当且仅当两变量独立时取值为零,当且仅当两变量之间存在完全函数关系时取值为一。我们利用这一性质,将查特吉相关系数用于评估非线性模型的预测能力,并进一步将其从单变量情形推广至多变量与高维情形。在高维设定下,我们完整刻画了基于查特吉相关系数的独立性检验的功效退化现象,给出了其渐近性质的理论结果。本报告的结果为在高维数据中选择合适的相关性度量、评估模型预测能力提供了理论依据与方法指导。
报告人简介:朱利平,中国人民大学教授、博士生导师,国家级人才,享受国务院特殊津贴专家,兼任人民教育出版社普通高中教材《数学》分册联合主编,《中国大百科全书》统计学科编委会委员和数据科学分卷主编等。先后受邀担任国际统计学领域顶级学术期刊《美国统计学会会刊》和《统计年刊》、国际权威学术期刊《中华统计学》和《多元分析期刊》等副主编,以及国内统计学领域顶级学术期刊《中国科学·数学》(中、英文版)和《系统科学与数学》(中、英文版)等专业期刊青年编委、编委和副主编等。长期从事高维度和非线性大数据统计分析研究工作,多项研究成果被收录进多本国内外出版的重要教材和专著中。
第五场
报告承办单位: 数学与统计学院
报告题目: Simplex Frank-Wolfe: Linear Convergence and Its Numerical Efficiency
报告人姓名: 戚厚铎
报告人所在单位: 香港理工大学
报告人职称: 教授、博士生导师
报告时间: 2026 年 11 月 9 日 星期一 上午 10:30
报告地点: 理科楼 A212
报告摘要: We investigate variants of the Frank-Wolfe (FW) algorithm for smoothing and strongly convex optimization over polyhedral sets, with the goal of designing algorithms that achieve linear convergence while minimizing per-iteration complexity as much as possible. Starting from the simple yet fundamental unit simplex, and based on geometrically intuitive motivations, we introduce a novel oracle called Simplex Linear Minimization Oracle (SLMO), which can be implemented with the same complexity as the standard FW oracle. We then present two FW variants based on SLMO: Simplex Frank-Wolfe and the refined Simplex Frank-Wolfe (rSFW). Finally, we generalize the entire framework from the unit simplex to arbitrary polytopes. Furthermore, the refinement step in rSFW can accommodate any existing FW strategies such as the well-known away-step and pairwise-step, leading to outstanding numerical performance.
报告人简介:戚厚铎,香港理工大学数据科学及人工智能学系与应用数学系联合聘任教授,曾任英国南安普顿大学优化方向教授。他于 1996 年获得中国科学院博士学位,研究兴趣包括数据科学中的矩阵优化、稀疏优化以及投资组合优化等。他曾获澳大利亚研究理事会伊丽莎白二世女王研究员基金(QEII Fellow,2003 年),并于 2019 年获英国国家数据科学研究所艾伦·图灵研究所的图灵研究员(Turing Fellow)称号。现任Asia-Pacific Journal of Operational Research 优化领域编辑,以及 Mathematical Programming Computation, Journal of Operations Research Society of China,Computational Optimization and Applications, and Numerical Functional Analysis and Optimization 等期刊的副主编。担任英国工程与物理科学研究理事会(EPSRC)评审学院成员,并担任欧洲多个研究理事会重大资助项目的评审委员会成员。