学术演讲：2023年世界顶尖科学家协会奖“智能科学或数学奖”获奖者讲堂

发布日期 2023年11月23日

点击此处跳转世界顶尖科学家论坛官网观看视频回放。

在刚刚于上海结束的第六届世界顶尖科学家论坛期间，2023年世界顶尖科学家协会奖“智能科学或数学奖”的两位获奖者相继发表了学术演讲。世界顶尖科学家协会奖获奖者讲堂是顶科协奖年度系列活动的一部分。

主持人：迈克尔·I·乔丹（Michael I. Jordan），2023年世界顶尖科学家协会奖“智能科学或数学奖”遴选委员会主席，加州大学伯克利分校电子工程与计算机科学系、统计学系杰出冠名教授

演讲嘉宾：2023年世界顶尖科学家协会奖“智能科学或数学奖”获奖者

尤里·涅斯捷罗夫（Yurii Nesterov）
比利时法语鲁汶大学运筹学与计量经济学研究中心、数学工程系名誉教授、高级科学研究员
演讲标题：Optimization, the Philosophical Background of Artificial Intelligence
演讲梗概：
We discuss new challenges in the modern Science, created by Artificial Intelligence (AI). Indeed, AI requires a system of new sciences, mainly based on computational models. Their development has already started by the progress in Computational Mathematics. In this new reality, Optimization plays an important role, helping the other fields with finding tractable models and efficient methods, and significantly increasing their predictive power. We support our conclusions by several examples of efficient optimization schemes related to human activity.

阿尔卡迪·涅米罗夫斯基（Arkadi Nemirovski）
美国佐治亚理工学院工业与系统工程学院讲席教授
演讲标题：Topics in Convex Optimization
演讲梗概：
We present an informal "executive summary" of "well-structured" convex optimization problems and their "structure-revealing" conic representations, primarily polyhedral, conic quadratic, and semidefinite ones. As a matter of fact, basically all convex problems arising in applications admit representations of this type, making the problems amenable to theoretically (and to some extent also practically) efficient solution algorithms capable of approximating to whatever high accuracy globally optimal solutions in a reasonable time. Another advantage of conic representations is the existence of a fully algorithmic duality theory, entirely similar to Linear Programming duality. Conic duality is indispensable in algorithmic design and analysis, on the one hand, and offers powerful tools for instructive processing convex optimization models "on paper," on the other hand. In the lecture, we entirely omit algorithmic issues (too technical for informal presentation) and focus on Conic Programming duality, revealing its surprisingly simply-looking geometry and briefly outlining several applications, including those in Truss Topology Design, Robust Optimization, synthesis of robust linear controllers, and Signal Estimation.