CAREER: Deep Learning Based Scientific Computing: Mathematical Theory and Algorithms
职业:基于深度学习的科学计算:数学理论与算法
基本信息
- 批准号:2244988
- 负责人:
- 金额:$ 42.56万
- 依托单位:
- 依托单位国家:美国
- 项目类别:Continuing Grant
- 财政年份:2022
- 资助国家:美国
- 起止时间:2022-07-01 至 2025-06-30
- 项目状态:未结题
- 来源:
- 关键词:
项目摘要
Deep learning has demonstrated remarkable, high fidelity performance on computer vision and natural language processing tasks that revolutionize manufacturing and social life. Recent applications of deep learning in scientific problems have also advanced scientific discovery via computational chemistry, materials science, medicine, immunology, climate sciences, etc. Understanding the mathematical principles of deep learning algorithms is crucial to validating and improving these algorithms, and will allow scientists and engineers to obtain more reliable predictions and perform a better risk assessment. The research goal is to develop a systematic deep learning analysis serving as the theoretical foundation of numerous scientific problems based on deep learning; cutting-edge algorithms for the efficient solutions of high-dimensional and highly nonlinear partial differential equations arising in various application domains will also be proposed with a theoretical guarantee. The proposed deep learning-based algorithms for high-dimensional and highly nonlinear problems will be expected to greatly advance the state-of-the-art simulations of complex physical systems arising in many fields in science and engineering. The theoretical challenges of deep learning are largely due to the highly non-linear nature of deep neural networks (DNNs). As a function parametrization tool formulated as compositions of non-linear functions, DNNs are highly non-linear and require advanced mathematics to fully understand. Therefore, there is a critical need for new advances in mathematics for a better understanding of DNNs. The theoretical part of this project mainly focuses on the approximation and generalization capacity of DNNs. The central questions to be answered are whether DNN approximation conquers or lessens the curse of dimensionality, what is the optimal approximation rate of various function classes, and how to characterize the Rademacher complexity of various DNNs trained with state-of-the-art empirical regularization methods aiming at optimal generalization error bound. The computational part of this project concentrates on solving high dimensional and highly oscillatory partial differential equations. The specific approach of this project is to propose hybrid algorithms that combine the advantage of deep learning algorithms and traditional numerical techniques for more efficient computation and higher accuracy. The key idea is to treat deep learning solvers as a preconditioner of traditional numerical algorithms. The algorithms designed in the project will also be implemented in deep learning packages for numerical PDEs and made publicly available. Research outcomes of this project will be disseminated through conferences, publications (journal papers and textbooks), and new mathematical deep learning courses to a broad audience, especially for the next generation of computational scientists.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
深度学习在计算机视觉和自然语言处理任务中表现出显著的高保真性能,这些任务给制造业和社会生活带来了革命性的变化。近年来,深度学习在科学问题中的应用也推动了计算化学、材料科学、医学、免疫学、气候科学等科学发现的发展。理解深度学习算法的数学原理对于验证和改进这些算法至关重要,将使科学家和工程师获得更可靠的预测和更好的风险评估。研究的目标是开发一个系统的深度学习分析,作为基于深度学习的众多科学问题的理论基础;还将提出解决各种应用领域中出现的高维和高度非线性偏微分方程组的有效解的前沿算法,并提供理论保证。所提出的基于深度学习的高维和高度非线性问题的算法将极大地促进科学和工程中许多领域出现的复杂物理系统的最新模拟。深度学习的理论挑战很大程度上源于深度神经网络(DNN)的高度非线性。作为一种由非线性函数组合而成的函数参数化工具,DNN是高度非线性的,需要高等数学才能完全理解。因此,迫切需要在数学方面取得新的进展,以便更好地理解DNN。本项目的理论部分主要关注DNN的逼近和泛化能力。需要回答的核心问题是DNN逼近是否克服或减少了维度灾难,各种函数类的最优逼近速度是多少,以及如何表征以最优泛化误差界为目标的各种经验正则化方法训练的DNN的Rademacher复杂性。这个项目的计算部分集中在求解高维和高度振荡的偏微分方程组。该项目的具体方法是提出结合深度学习算法和传统数值技术的优点的混合算法,以实现更高效的计算和更高的精度。其关键思想是将深度学习求解器作为传统数值算法的预条件。该项目中设计的算法还将在数值PDE的深度学习包中实现,并公开可用。该项目的研究成果将通过会议、出版物(期刊论文和教科书)和新的数学深度学习课程向广大受众传播,特别是针对下一代计算科学家。该奖项反映了NSF的法定使命,并通过使用基金会的智力优势和更广泛的影响审查标准进行评估,被认为值得支持。
项目成果
期刊论文数量(12)
专著数量(0)
科研奖励数量(0)
会议论文数量(0)
专利数量(0)
From Optimization Dynamics to Generalization Bounds via {\L}ojasiewicz Gradient Inequality
通过 {L}ojasiewicz 梯度不等式从优化动态到泛化界限
- DOI:
- 发表时间:2022
- 期刊:
- 影响因子:0
- 作者:Fusheng Liu;Haizhao Yang;Soufiane Hayou;Qianxiao Li
- 通讯作者:Qianxiao Li
Stationary Density Estimation of Itô Diffusions Using Deep Learning
使用深度学习的 Ità 扩散的稳态密度估计
- DOI:10.1137/21m1445363
- 发表时间:2023
- 期刊:
- 影响因子:2.9
- 作者:Gu, Yiqi;Harlim, John;Liang, Senwei;Yang, Haizhao
- 通讯作者:Yang, Haizhao
Linear-scaling selected inversion based on hierarchical interpolative factorization for self Green's function for modified Poisson-Boltzmann equation in two dimensions
- DOI:10.1016/j.jcp.2021.110893
- 发表时间:2021-05
- 期刊:
- 影响因子:0
- 作者:Y. Tu;Qiyuan Pang;Haizhao Yang;Zhenli Xu
- 通讯作者:Y. Tu;Qiyuan Pang;Haizhao Yang;Zhenli Xu
PiPs: A Kernel-based Optimization Scheme for Analyzing Non-Stationary 1D Signals
- DOI:10.1016/j.acha.2023.04.002
- 发表时间:2018-05
- 期刊:
- 影响因子:2.5
- 作者:Jieren Xu;Yitong Li;Haizhao Yang;D. Dunson;I. Daubechies
- 通讯作者:Jieren Xu;Yitong Li;Haizhao Yang;D. Dunson;I. Daubechies
Friedrichs Learning: Weak Solutions of Partial Differential Equations via Deep Learning
- DOI:10.2139/ssrn.3964424
- 发表时间:2020-12
- 期刊:
- 影响因子:0
- 作者:Fan Chen;J. Huang;Chunmei Wang;Haizhao Yang
- 通讯作者:Fan Chen;J. Huang;Chunmei Wang;Haizhao Yang
{{
item.title }}
{{ item.translation_title }}
- DOI:
{{ item.doi }} - 发表时间:
{{ item.publish_year }} - 期刊:
- 影响因子:{{ item.factor }}
- 作者:
{{ item.authors }} - 通讯作者:
{{ item.author }}
数据更新时间:{{ journalArticles.updateTime }}
{{ item.title }}
- 作者:
{{ item.author }}
数据更新时间:{{ monograph.updateTime }}
{{ item.title }}
- 作者:
{{ item.author }}
数据更新时间:{{ sciAawards.updateTime }}
{{ item.title }}
- 作者:
{{ item.author }}
数据更新时间:{{ conferencePapers.updateTime }}
{{ item.title }}
- 作者:
{{ item.author }}
数据更新时间:{{ patent.updateTime }}
Haizhao Yang其他文献
Phase-Space Sketching for Crystal Image Analysis Based on Synchrosqueezed Transforms
基于同步压缩变换的晶体图像分析相空间草图
- DOI:
10.1137/17m1129441 - 发表时间:
2017 - 期刊:
- 影响因子:0
- 作者:
Jianfeng Lu;Haizhao Yang - 通讯作者:
Haizhao Yang
Multiresolution mode decomposition for adaptive time series analysis
- DOI:
10.1016/j.acha.2019.09.006 - 发表时间:
2017-09 - 期刊:
- 影响因子:2.5
- 作者:
Haizhao Yang - 通讯作者:
Haizhao Yang
Let Data Talk: Data-regularized Operator Learning Theory for Inverse Problems
让数据说话:反问题的数据正则化算子学习理论
- DOI:
10.48550/arxiv.2310.09854 - 发表时间:
2023 - 期刊:
- 影响因子:0
- 作者:
Ke Chen;Chunmei Wang;Haizhao Yang - 通讯作者:
Haizhao Yang
Statistical analysis of synchrosqueezed transforms
- DOI:
10.1016/j.acha.2017.01.001 - 发表时间:
2014-10 - 期刊:
- 影响因子:2.5
- 作者:
Haizhao Yang - 通讯作者:
Haizhao Yang
Oscillatory data analysis and fast algorithms for integral operators
- DOI:
- 发表时间:
2015-06 - 期刊:
- 影响因子:0
- 作者:
Haizhao Yang - 通讯作者:
Haizhao Yang
Haizhao Yang的其他文献
{{
item.title }}
{{ item.translation_title }}
- DOI:
{{ item.doi }} - 发表时间:
{{ item.publish_year }} - 期刊:
- 影响因子:{{ item.factor }}
- 作者:
{{ item.authors }} - 通讯作者:
{{ item.author }}
{{ truncateString('Haizhao Yang', 18)}}的其他基金
Collaborative Research: Friedrichs Learning: Mathematical Foundation and Applications
合作研究:弗里德里希学习:数学基础与应用
- 批准号:
2206333 - 财政年份:2022
- 资助金额:
$ 42.56万 - 项目类别:
Standard Grant
CAREER: Deep Learning Based Scientific Computing: Mathematical Theory and Algorithms
职业:基于深度学习的科学计算:数学理论与算法
- 批准号:
1945029 - 财政年份:2020
- 资助金额:
$ 42.56万 - 项目类别:
Continuing Grant
相似国自然基金
基于Deep Unrolling的高分辨近红外二区荧光分子断层成像方法研究
- 批准号:12271434
- 批准年份:2022
- 资助金额:46 万元
- 项目类别:面上项目
基于深度森林(Deep Forest)模型的表面增强拉曼光谱分析方法研究
- 批准号:2020A151501709
- 批准年份:2020
- 资助金额:10.0 万元
- 项目类别:省市级项目
面向Deep Web的数据整合关键技术研究
- 批准号:61872168
- 批准年份:2018
- 资助金额:62.0 万元
- 项目类别:面上项目
基于Deep-learning的三江源区冰川监测动态识别技术研究
- 批准号:51769027
- 批准年份:2017
- 资助金额:38.0 万元
- 项目类别:地区科学基金项目
具有时序处理能力的Spiking-Deep Learning(脉冲深度学习)方法研究
- 批准号:61573081
- 批准年份:2015
- 资助金额:64.0 万元
- 项目类别:面上项目
基于语义计算的海量Deep Web知识探索机制研究
- 批准号:61272411
- 批准年份:2012
- 资助金额:80.0 万元
- 项目类别:面上项目
Deep Web数据集成查询结果抽取与整合关键技术研究
- 批准号:61100167
- 批准年份:2011
- 资助金额:20.0 万元
- 项目类别:青年科学基金项目
面向Deep Web的大规模知识库自动构建方法研究
- 批准号:61170020
- 批准年份:2011
- 资助金额:57.0 万元
- 项目类别:面上项目
Deep Web敏感聚合信息保护方法研究
- 批准号:61003054
- 批准年份:2010
- 资助金额:20.0 万元
- 项目类别:青年科学基金项目
基于逻辑强化学习的Deep Web模式匹配研究
- 批准号:61070122
- 批准年份:2010
- 资助金额:32.0 万元
- 项目类别:面上项目
相似海外基金
CAREER: Adaptive Deep Learning Systems Towards Edge Intelligence
职业:迈向边缘智能的自适应深度学习系统
- 批准号:
2338512 - 财政年份:2024
- 资助金额:
$ 42.56万 - 项目类别:
Continuing Grant
CAREER: Theoretical foundations for deep learning and large-scale AI models
职业:深度学习和大规模人工智能模型的理论基础
- 批准号:
2339904 - 财政年份:2024
- 资助金额:
$ 42.56万 - 项目类别:
Continuing Grant
CAREER: Protecting Deep Learning Systems against Hardware-Oriented Vulnerabilities
职业:保护深度学习系统免受面向硬件的漏洞的影响
- 批准号:
2426299 - 财政年份:2024
- 资助金额:
$ 42.56万 - 项目类别:
Continuing Grant
CAREER: Physics-Informed Deep Learning for Understanding Earthquake Slip Complexity
职业:基于物理的深度学习用于理解地震滑动的复杂性
- 批准号:
2339996 - 财政年份:2024
- 资助金额:
$ 42.56万 - 项目类别:
Continuing Grant
CAREER: Enhancing Temperature Visualization in Boiling Fluid over Finned Surfaces using Deep Learning-Enhanced Laser-Induced Fluorescence
职业:使用深度学习增强激光诱导荧光增强翅片表面沸腾流体的温度可视化
- 批准号:
2337973 - 财政年份:2024
- 资助金额:
$ 42.56万 - 项目类别:
Continuing Grant
CAREER: Accelerating Scientific Discovery via Deep Learning with Strong Physics Inductive Biases
职业:通过具有强物理归纳偏差的深度学习加速科学发现
- 批准号:
2338909 - 财政年份:2024
- 资助金额:
$ 42.56万 - 项目类别:
Continuing Grant
CAREER: Geometric Deep Learning to Facilitate Algorithmic and Scientific Advances in Therapeutics
职业:几何深度学习促进治疗学的算法和科学进步
- 批准号:
2339524 - 财政年份:2024
- 资助金额:
$ 42.56万 - 项目类别:
Continuing Grant
CAREER: Manufacturing USA: Deep Learning to Understand Fatigue Performance and Processing Relationship of Complex Parts by Additive Manufacturing for High-consequence Applications
职业:美国制造:通过深度学习了解复杂零件的疲劳性能和加工关系,通过增材制造实现高后果应用
- 批准号:
2239307 - 财政年份:2023
- 资助金额:
$ 42.56万 - 项目类别:
Standard Grant
CAREER: Accelerating Probabilistic Predictions of Sea-level Rise with Deep Learning
职业:利用深度学习加速海平面上升的概率预测
- 批准号:
2238316 - 财政年份:2023
- 资助金额:
$ 42.56万 - 项目类别:
Standard Grant
CAREER: Efficient, Dynamic, Robust, and On-Device Continual Deep Learning with Non-Volatile Memory based In-Memory Computing System
职业:使用基于非易失性内存的内存计算系统进行高效、动态、鲁棒、设备上持续深度学习
- 批准号:
2342726 - 财政年份:2023
- 资助金额:
$ 42.56万 - 项目类别:
Continuing Grant