报告时间:2026年11月19日19:30-21:40
报告地点:腾讯会议:306-883-207
校内联系人:韩月才 [email protected]
报告一: On an Iwaniec--Kovalev--Onninen Conjecture for harmonic quasiconformal annulus mappings
报告人:刘劲松研究员,中国科金沙娱乐城
报告摘要:In this talk, I would like to report a recent joint work regarding to the Iwaniec--Kovalev--Onninen Conjecture. In fact, we disprove the upper Nitsche–Gr\"otzsch conjecture for unrestricted non-radial harmonic quasiconformal mappings between circular annuli. For every pair in which the target annulus is thicker than the source, we construct a smooth, orientation-preserving, non-radial harmonic diffeomorphism whose maximal dilatation is strictly below the conjectured radial threshold. The key innovation is a high-frequency boundary reparametrization that reduces the principal Fourier-mode distortion, while the compensating high-frequency modes decay exponentially toward the inner boundary. We also identify energy, spectral, and low-frequency conditions under which the original estimate remains valid. Its minimal-surface interpretation produces vertical-periodic counterexamples and motivates a new non-radial extremal problem. This is a joint work with David Kalaj and Jian-Feng Zhu.
简介:刘劲松,中国科金沙娱乐城
数学与系统科学研究院研究员,主要研究兴趣是Teichmuller空间,主持基金委的杰出青年科学基金和重点项目,其研究论文发表在Invent. Math.、Math. Ann.、Adv. Math.、Trans. Amer. Math. Soc.、Sci. China Math.等国际著名数学期刊.
报告二: Hyperbolic approximation and the estimates of dimensions
报告人:曹永罗教授,苏州大学
报告摘要: In this talk, we report some recent results about hyperbolic approximation, and then we use this approximation to give the estimates of dimensions for nonconformal repellers. We also report some results about the multifractal analysis of Birkhoff average for almost self-affine systems.
简介:苏州大学教授、博士生导师,主要从事微分动力系统遍历理论以及动力系统维数理论的研究。2011年获国家自然科学杰出青年基金,2013年获江苏省教学成果奖一等奖,2023年获教育部自然科学一等奖;曾主持国家自然科学基金重大项目课题、国家科技部重点研发计划课题,以及多项国家自然科学基金和973项目基金。在GAFA、CMP、Trans. AMS.、IMRN、Math.Z、ETDS、JDE、JFA、Nonlinearity等国际权威期刊上发表论文60余篇,目前担任期刊英国Dynamical Systems:An International Journal,以及DCDS(S)期刊编委。
报告三: The Fairness Hierarchy: A viewpoint from causal inference
报告人:李长城教授,大连理工大学
报告摘要: Fairness in machine learning predictions has received increasing attention in recent years.In this work, we develop a causal inference--based framework for generating predictions that satisfy different levels of fairness. Unlike existing approaches that enforce fairness via constraints on predictive models or objectives, we focus on counterfactual prediction semantics induced by path-specific interventions.We characterize fairness levels through causal paths, where sensitive information, possibly arising from multiple sources, is selectively blocked or allowed to propagate to the prediction outcome $Y$ along specified routes. Based on a learned causal graph, we construct a probabilistic predictive distribution for $Y$ using a structural causal model (SCM), and perform counterfactual predictive inference by intervening on designated causal paths from sensitive attributes to the outcome. By flexibly selecting which causal paths are subject to counterfactual intervention, our framework provides a unified and interpretable hierarchy of fairness notions, generalizing standard path-specific causal semantics. We illustrate the proposed framework through empirical experiments to demonstrate how different fairness levels can be operationalized in practice.
简介:李长城,大连理工大学数学科学金沙娱乐城
教授。研究兴趣主要包括高维统计推断及高维因果推断。在高维统计的理论、应用以及计算方面进行了一系列创新性的研究,在一流学术期刊JASA、AOS、JOE、AOAS、Statistica Sinica等发表文章多篇,入选国家级、省级青年人才计划。