本期编辑:
周涛
中国科学院数学与系统科学研究院
tzhou@lsec.cc.ac.cn
周知
香港理工大学应用数学系
zhizhou@polyu.edu.hk
内容提要:
====================================================================================
2022年度中国工业与应用数学学会专业委员会工作会议在线上召开
来源:中国工业与应用数学学会网站
2023年1月6日,2022年度中国工业与应用数学学会(以下简称“学会”)专业委员会工作会议在
线上召开。学会理事长张平文院士、秘书长张波研究员、专业委员会管理与地方学会联络委员会
(以下简称“管理委员会”)委员、各专业委员会负责人出席本次会议。学会副理事长、管理委员
会主任黄云清教授主持本次工作会议。
学会理事长张平文院士首先致辞。他代表学会感谢了各专业委员会长久以来的付出与努力,对各
专业委员会能克服疫情影响仍保持很强的学术活跃度表达了高度的称赞。他指出,今年将要在日
本召开的ICIAM 2023大会是学会及各专业委员会扩大国际学术影响力的难得契机,他鼓励各专
业委员会可以借助这次盛会多多发出中国声音。随着国内应用数学的发展进入一个崭新的时代,
他建议各专业委员会可以在推动应用数学的落地、学科的交叉、基础算法的研究等方面抓住机会,
寻求更加长足的进步。此外,他希望专业委员会在得到极大发展的同时,也要兼顾加强自身建设,
积极思考和推进工作的专职化与专业度。最后,理事长给所有与会人员送出了新春祝福,对学会、
专业委员会、应用数学未来的繁荣发展表达了美好的祝愿。
接下来,进入专业委员会汇报阶段,筹备中的1个专业委员会和已正式成立的21个专业委员会依
次作了年度工作总结。每个专业委员会都从组织架构、学术活动开展情况、科普活动开展情况等
方面进行了汇报,并提出了2023年的总体工作计划。管理委员会在听取各专业委员会工作汇报的
同时,对其年度工作开展情况进行了现场提问与评估打分。
所有专业委员会汇报完毕后,学会副理事长黄云清教授作了总结发言。他肯定了各专业委员会近
些年在管理规范化方面做出的努力与巨大进步,对各专业委员会负责人所做的贡献表示了由衷的
感谢。他指出,许多专业委员会能够跟据自身情况与学科特点开展特色活动,尤其是各类研讨会
和科普活动丰富多彩、形式多样,值得相互学习、继续发扬。随着国家疫情防控战略的进一步放
开,他鼓励各专业委员会在疫情平稳的前提下,未来可以多尝试组织线下活动,以营造更好的学
术交流氛围,取得更好的学术活动效果。最后,黄教授希望各专业委员会应避免盲目扩张,要逐
步重视起发展的质量,提高会员的凝聚力,同时加强专业委员会横向之间的交流合作、联合攻关,
引导会员开展有组织的科研,力争理论研究与解决实际问题齐头并进,助力更多的应用数学成果
落地。
本次会议的最后,管理委员会进行了合议。合议环节中,管理委员会委员们就本次汇报中出现的
共性问题展开了讨论,研究了工作中遇到的若干问题,同时结合本年度的汇报情况对今后的考核
标准进行了讨论与优化。
本次工作会议对2022年度各专业委员会所做的工作进行了评估考核,并对工作中遇到的问题进
行了针对性研究。通过本次会议,各专业委员会也得以相互学习、交流经验、取长补短,对专业
委员会的良性发展、加强合作具有重要意义。2022年是新冠病毒传播以来形势最为严峻的一年,
全国各地封控不断,对举办学术活动造成了巨大影响,但各专业委员会克服了各种困难,通过灵
活的形式依旧保持了很高的学术活跃度,难能可贵;2022年也是新冠病毒传播的转折之年,相
信在已经到来的2023年,各专业委员会所举办的活动效果及自身的发展建设都将向前再迈进一
大步。
学会专业委员会管理与地方学会联络委员会供稿
====================================================================================
来源:中国工业与应用数学学会网站
URL: https://www.csiam.org.cn/home/article/detail/id/1872.html
为鼓励更多的科研项目成果转化落地,促进应用数学及相关工业发展,中国工业与应用数学学会自
2021年起评选认证CSIAM应用数学落地成果,历年已获得认证的落地成果名单请见
https://www.csiam.org.cn/home/article/detail/id/1456.html。
详情请见网站:http://csiam.math2industry.org.cn/amat/。
2023年CSIAM应用数学落地成果征集工作现已启动。
一、申请认证项目应具备以下资格:
1、科研项目成果转化落地;
2、主要成果在中国完成。
二、认证采用推荐制度,由以下两种方式推荐产生:
1、个人推荐需符合下列条件之一(须两位推荐人):
(1)中国科学院或中国工程院院士;
(2)国内有重大影响力(中国500强)企业负责研发创新合作的副总裁及以上职位;
(3)CSIAM会士、CSIAM常务理事、CSIAM专业委员会主任;
(4)省级工业与应用数学学会理事长。
2、单位推荐:CSIAM团体会员。
*注:因CSIAM团体会员名单每月实时更新,请以2022年12月《简讯》中的名单(或查看附件中
的团体会员名单,二者一致)为准。
每位推荐人和推荐单位每年限推荐一项成果。
三、候选落地成果需提交以下材料:
不超过1000字的项目简介;
不超过10分钟的项目视频介绍;
2份个人推荐信或1份单位推荐信。
注:视频中请清楚描述项目问题背景、项目所涉及的数学问题、解决方法、以及成果应用后所产
生的效益/性能/影响等。
四、请于2023年3月5日(含)前将申请材料上传提交至网站:
https://math2industry.org.cn/certification。
五、联系方式:
联系人:张梦思
电 话:157-3886-8281
邮 箱:mszhang@csiam.org.cn
中国工业与应用数学学会
2023年1月5日
====================================================================================
来源:中国数学会网站
URL: http://www.cms.org.cn/Home/notices/notices_details/id/987.html
尊敬的各位老师:
大家好!为了更好地推动各学科的学术交流,中国数学会将对2023年拟将召开的部分重要学术会
议给予适当支持,如您单位有重要的学术会议需要中国数学会的支持,请提交以下材料:
1.会议资助申请表WORD版及PDF版(PDF版需在相应位置盖章签字,模板见附件)。
2.会议日程WORD版。
3.会议邀请报告人信息(包含姓名、单位等)WORD版。
4.科普活动计划,WORD版。
以上材料请于2023年2月28日前发送至中国数学会办公室,过时将不予受理。
注意事项:
1.仅支持学术会议,针对同一个机构同一年举办的会议原则上只支持1个。
2.会议名称原则上不能带有“全国”“中国”等字眼。
3.申请中国数学会资助的会议需要开展至少一次的科普活动,并录制科普活动视频,以便在中国
数学会B站进行宣传。会后将科普活动视频及会议纪要(需记录科普情况)反馈至数学会办公室。
4.学术带头人最好由国际国内的著名专家担任。
5.国际会议应包含数量较多的国外专家学者,材料提交尽量齐全,需要提交程序委员会名单。系
列国际会议需提交会议前三届会议召开情况的文字介绍,比如历届会议地点,重要参会人员等。
感谢您的支持!
中国数学会办公室
2023年1月28日
====================================================================================
来源:中国数学会网站
URL: http://www.cms.org.cn/Home/notices/notices_details/id/992.html
Thematic Session Proposals
Call for Thematic Session Proposals at the China-Brazil Joint
Mathematical Meeting – Foz do Iguacu, Brazil.
Proposals of thematic sessions at China-Brazil Joint Mathematical
Meeting are welcomed by the Organizing Committee. Early submission
of proposals is encouraged: good proposals will be approved on a
regular basis before the deadline, so that session speakers may
be invited with enough time to make travel and funding arrangements.
A proposal should include :
the names, affiliations and contact information (including email addresses)
of all the organizers, with one organizer designated as “contact organizer”,
a title and a brief presentation of the topic and scope (up to one page),
the list of speakers including the tentative title and abstract of each talk.
Each special session will consist from 8 to 12 talks of 30 minutes
distributed over slots of at most 4 talks each.
The list of speakers must include mathematicians from China and Brazil.
Preference will be given to proposals whose list of suggested speakers
represents diversity in all aspects.
Proposals must be sent to: brazilchina2023@gmail.com and cc to the Chinese
Mathematical Society: cms@math.ac.cn.
More information can be found at:
https://sbm.org.br/jointmeeting-china/
or http://www.cms.org.cn/Home/notices/notices_details/id/990.html
Important dates
Call for thematic sessions: Oct. 31st, 2022
Session submission: extended deadline, Feb., 15th, 2023
Session acceptance notice: Jan. 30th — Feb. 28th, 2023
Abstract/Contributed talk submission: Feb., 1st — Mar., 1st, 2023
Acceptance notice: Apr., 1st, 2023
Early registration: Mar, 1st — Jun., 1st, 2023
Registration: Jul., 1st, 2023
Conference: Jul., 17 — 21, 2023
====================================================================================
来源:中国数学会网站
URL:http://www.cms.org.cn/Home/notices/notices_details/id/996.html
中国数学会2022年学术年会将于2023年2月18日—22日(18日报到,22日离会)在湖北省武汉
市举行。这是中国数学工作者一年一次的学术盛会,开幕式上将颁发中国数学会第十六届华罗庚
数学奖、第十九届陈省身数学奖和第十六届钟家庆数学奖。会上将邀请叶向东、张平、唐梓洲和
单芃等4位数学家作大会报告,邀请百余位数学家在代数与数论、几何与拓扑、常微动力系统、偏
微分方程、实分析和复分析、计算数学、概率和统计、运筹与控制、组合与计算机数学、数学史
与数学教育等10个领域作分组报告,还将邀请部分院士和专家在武汉地区大中院校作科普报告。
会议期间还将召开数学文化与传播论坛、中学生创新人才培养论坛等专业论坛,同时召开一些数
学专业领域的卫星会议。
会议注册和缴费以及酒店预定系统均已开通,敬请参会代表尽早按网页说明(http://ssci.whut.edu.cn/cms2022/)
尽快完成网上注册、并预订酒店房间。
本次会议由于疫情影响延期举办,已注册和缴费的人员无需再次注册和缴费。已提前缴费但不能
参会的人员请在2023年2月16日前发邮件至cms@math.ac.cn和songt@whut.edu.cn申请会
议费退款。具体缴费与开发票信息请参考《中国数学会缴费平台操作说明》。
具体信息参见
http://www.cms.org.cn/ueditor/php/upload/file/20230130/1675064069210659.pdf
====================================================================================
【会议信息】The 10th Heidelberg Laureate Forum, 2023
URL: https://www.heidelberg-laureate-forum.org/forum/10th-hlf-2023.html
The 10th Heidelberg Laureate Forum will take place from September 24 to 29
and brings together some of the brightest minds in mathematics and computer
science for an unrestrained, interdisciplinary exchange. During the weeklong
conference, young researchers and other participants have the opportunity
to connect with scientific pioneers and learn how the laureates made it to
the top of their fields. Laureate lectures and discussions, plus various
interactive program elements are some of the Forum's fundamental elements.
This compelling networking event combines scientific, social and outreach
activities in a unique atmosphere, sustained by comprehensive exchange and
scientific inspiration.
Young researchers can apply to attend the 10th HLF from Friday, November 11,
2022 until Saturday, February 11, 2023.
Journalists are encouraged to cover the forum and have the opportunity to
apply for a limited numbers of travel grants from February 20 to April 30.
Both groups can apply here: https://application.heidelberg-laureate-forum.org
For more information on the young researcher applications process, please
check out our FAQ.
Updates and detailed information will be made available on the website as
the program continues to materialize.====================================================================================
URL: https://applmath.cjoe.ac.cn/jweb_yysxxb/CN/current
时标上右端函数为两项和的集值微分方程解的平方收敛性
王培光, 吴曦冉
基于高频数据的GARCH模型拟极大指数似然估计
李莉丽, 张兴发, 邓春亮, 李元
一类具有季节交替的n维连续时间Leslie/Gower竞争模型
陈梅香, 谢溪庄
高阶非线性薛定谔方程的可积边界条件
王中园, 张成
一类具非标准增长条件和非强制项的抛物方程弱解的存在性
李仲庆
解一类函数极值问题的动约束同伦算法
商玉凤, 刘庆怀
时变参数不确定广义系统的有限时间预见控制器设计
李丽, 卢延荣
指数谱负Lévy过程下的清算风险
李鑫, 蒋峰
基于空间异质反应扩散HIV感染模型的最优治疗策略
吴鹏, 赵洪涌
独立增量过程下亚式期权的方差最优对冲策略
贾兆丽, 杨舒荃, 吴霍俊
====================================================================================
【期刊信息】Acta Mathematicae Applicatae Sinica (English Series), Volume 39, Issue 1, 2023
URL: https://applmath.cjoe.ac.cn/jweb_yysxxb_en/EN/current
Preface: Hyperbolic System of Conservation Laws and Related Topics
Fei-Min HUANG
On Subsonic and Subsonic-Sonic Flows with General Conservatives Force in Exterior Domains
Xumin GU, Tian-Yi WANG
Smooth Solution of Multi-dimensional Nonhomogeneous Conservation Law: Its Formula,
and Necessary and Sufficient Blowup Criterion
Time Decay Rate of Solutions Toward the Viscous Shock Waves under Periodic Perturbations
for the Scalar Conservation Law with Nonlinear Viscosity
Ye-chi LIU
Stability of a Composite Wave of Two Separate Strong Viscous Shock Waves for 1-D
Isentropic Navier-Stokes System
Lin CHANG
Boundary Layer Solution of the Boltzmann Equation for Specular Boundary Condition
Fei-min HUANG, Zai-hong JIANG, Yong WANG
The Time Asymptotic Expansion of the Bipolar Hydrodynamic Model for Semiconductors
Xiao-chun WU
Global Weak Entropy Solution of Nonlinear Ideal Reaction Chromatography System and Applications
Jing ZHANG, Hong-xia LIU, Tao PAN
Global Smooth Solution to the Incompressible Navier-Stokes-Landau-Lifshitz Equations
Guang-wu WANG, You-de WANG
Incompressible Limit of the Compressible Q-tensor System of Liquid Crystals
Yi-xuan WANG
Sharp Condition for Global Existence of Supercritical Nonlinear Schrödinger Equation
with a Partial Confinement
====================================================================================
【期刊信息】Journal of Machine Learning, Volume 1, Number 4, 2022
发件人:Zhiqin Xu <xuzhiqin@sjtu.edu.cn>
Journal of Machine Learning (JML, jml.pub) is a new journal, published by
Global Science Press and sponsored by the Center for Machine Learning Research,
Peking University & AI for Science Institute, Beijing. Professor Weinan E
serves as the Editor-in-Chief together with managing editors Jiequn Han,
Arnulf Jentzen, Qianxiao Li, Lei Wang, Zhi-Qin John Xu, Linfeng Zhang.
JML publishes high quality research papers in all areas of machine learning (ML),
including innovative algorithms, theory, and applications in all areas. The
journal emphasizes a balanced coverage of both theory and application.
An introduction to the fourth issue of Journal of Machine Learning.
Title: A Mathematical Framework for Learning Probability Distributions
Authors: Hongkang Yang
DOI: 10.4208/jml.221202, J. Mach. Learn., 1 (2022), pp. 373-431.
The modeling of probability distributions is an important branch of machine
learning. It became popular in recent years thanks to the success of deep
generative models in difficult tasks such as image synthesis and text conversation.
Nevertheless, we still lack a theoretical understanding of the good performance
of distribution learning models. One mystery is the following paradox: it is
generally inevitable that the model suffers from memorization (converges to
the empirical distribution of the training samples) and thus becomes useless,
and yet in practice the trained model can generate new samples or estimate
the probability density over unseen samples. Meanwhile, the existing models
are so diverse that it has become overwhelming for practitioners and researchers
to get a clear picture of this fast-growing subject. This paper provides a
mathematical framework that unifies all the well-known models, so that they
can be systemically derived based on simple principles. This framework enables
our analysis of the theoretical mysteries of distribution learning, in particular,
the paradox between memorization and generalization. It is established that the
model during training enjoys implicit regularization, so that it approximates
the hidden target distribution before eventually turning towards the empirical
distribution. With early stopping, the generalization error escapes from the
curse of dimensionality and thus the model generalizes well.
Title: Approximation of Functionals by Neural Network Without Curse of Dimensionality
Authors: Yahong Yang & Yang Xiang
DOI: 10.4208/jml.221018, J. Mach. Learn., 1 (2022), pp. 342-372.
Learning functionals or operators by neural networks is nowadays widely used
in computational and applied mathematics. Compared with learning functions by
neural networks, an essential difference is that the input spaces of functionals
or operators are infinite dimensional space. Some recent works learnt functionals
or operators by reducing the input space into a finite dimensional space.
However, the curse of dimensionality always exists in this type of methods.
That is, in order to maintain the accuracy of an approximation, the number
of sample points grows exponentially with the increase of dimension.
In this paper, we establish a new method for the approximation of functionals
by neural networks without curse of dimensionality. Functionals, such as linear
functionals and energy functionals, have a wide range of important applications
in science and engineering fields. We define Fourier series of functionals and
the associated Barron spectral space of functionals, based on which our new
neural network approximation method is established. The parameters and the
network structure in our method only depend on the functional. The approximation
error of the neural network is $O(1/\sqrt{m})$ where $m$ is the size of the
network, which does not depend on the dimensionality.
------------------------------
End of CAM Digest
本期到此结束