CAREER: Computational Foundations of Modern Machine Learning

职业:现代机器学习的计算基础

基本信息

  • 批准号:
    2239265
  • 负责人:
  • 金额:
    $ 55万
  • 依托单位:
  • 依托单位国家:
    美国
  • 项目类别:
    Continuing Grant
  • 财政年份:
    2023
  • 资助国家:
    美国
  • 起止时间:
    2023-02-01 至 2028-01-31
  • 项目状态:
    未结题

项目摘要

Machine learning is poised to play a central role in various areas of society, including healthcare, transportation, education, and commerce. Despite this immense potential, there are significant gaps in the theoretical understanding of some of the most foundational aspects which are crucial for these modern applications. These applications pose complex requirements, such as memory or space constraints on the learning algorithm and robustness of the learned model to changes in the data distribution. Classical learning theory, however, mainly focuses on more traditionally examined metrics, such as the running time of the learning algorithm and the average error that the model obtains on a test set. Therefore, the goal of this project is to re-examine some of the foundations of learning theory and develop a theory that takes into account modern requirements such as memory efficiency and robustness. In doing this, the project will not only help build a rich algorithmic suite to meet these requirements in practice and hence significantly increase the scope of current applications; it will also develop insights that can guide the understanding of new theoretical angles on learning and computation.In more detail, the project’s objective is to understand the fundamental limits and trade-offs of what is achievable under these contemporary requirements and use this understanding to develop new algorithmic frameworks to meet the requirements. To achieve this, the project will examine if there are inherent trade-offs between the available memory and the best achievable convergence rate for a number of fundamental learning and optimization problems. The project aims to leverage these trade-offs to identify suitable classes of problems where it is possible to achieve the convergence rate of memory-intensive algorithms but with much less memory usage. Finally, the project aims to establish a principled framework that goes beyond the classical training/test paradigm to understand the generalization abilities of learned models, and to develop a toolbox that effectively addresses modern robustness demands. To aid in the translation of theoretical results to practical settings, open-source software will be developed, and algorithms will be evaluated on benchmark datasets. An educational plan is tightly integrated with the research objectives of the project, including outreach activities with high-school students and a collaboration with the Los Angeles County Office of Education.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.
机器学习有望在社会的各个领域发挥核心作用,包括医疗保健、交通、教育和商业。尽管有着巨大的潜力,但在对这些现代应用至关重要的一些最基本方面的理论理解方面仍存在重大差距。这些应用提出了复杂的要求,例如对学习算法的存储器或空间限制以及学习模型对数据分布变化的鲁棒性。 然而,经典学习理论主要关注传统上更受关注的指标,例如学习算法的运行时间和模型在测试集上获得的平均误差。因此,本项目的目标是重新审视学习理论的一些基础,并开发一种考虑到现代需求(如记忆效率和鲁棒性)的理论。在这样做的过程中,该项目不仅将帮助建立一个丰富的算法套件,以满足这些实际要求,从而大大增加目前的应用范围;它还将发展可以指导学习和计算的新理论角度的理解的见解。更详细地说,该项目的目标是了解基本的限制和贸易-在这些当代要求下可以实现的东西,并使用这种理解来开发新的算法框架来满足要求。为了实现这一目标,该项目将研究在一些基本学习和优化问题的可用内存和最佳可实现收敛速度之间是否存在固有的权衡。该项目旨在利用这些权衡来确定合适的问题类别,在这些问题中,可以实现内存密集型算法的收敛速度,但内存使用量要少得多。最后,该项目旨在建立一个超越经典训练/测试范式的原则性框架,以了解学习模型的泛化能力,并开发一个有效满足现代鲁棒性需求的工具箱。为了帮助将理论结果转化为实际情况,将开发开放源码软件,并在基准数据集上评估算法。教育计划与项目的研究目标紧密结合,包括与高中学生的外展活动以及与洛杉矶县教育办公室的合作。该奖项反映了NSF的法定使命,并通过使用基金会的智力价值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(2)
专著数量(0)
科研奖励数量(0)
会议论文数量(0)
专利数量(0)
NeuroSketch: Fast and Approximate Evaluation of Range Aggregate Queries with Neural Networks
Fairness in matching under uncertainty
不确定性下匹配的公平性
{{ 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 }}

Vatsal Sharan其他文献

On the Statistical Complexity of Sample Amplification
关于样本扩增的统计复杂性
  • DOI:
  • 发表时间:
    2022
  • 期刊:
  • 影响因子:
    0
  • 作者:
    Brian Axelrod;Shivam Garg;Yanjun Han;Vatsal Sharan;G. Valiant
  • 通讯作者:
    G. Valiant
Transformers Learn Higher-Order Optimization Methods for In-Context Learning: A Study with Linear Models
Transformers 学习用于上下文学习的高阶优化方法:线性模型的研究
  • DOI:
    10.48550/arxiv.2310.17086
  • 发表时间:
    2023
  • 期刊:
  • 影响因子:
    0
  • 作者:
    Deqing Fu;Tian;Robin Jia;Vatsal Sharan
  • 通讯作者:
    Vatsal Sharan
Understanding the Capabilities and Limitations of Neural Networks for Multi-task Learning
了解神经网络用于多任务学习的能力和局限性
  • DOI:
  • 发表时间:
    2019
  • 期刊:
  • 影响因子:
    0
  • 作者:
    Vatsal Sharan;Xin Wang;Brendan Juba;R. Panigrahy
  • 通讯作者:
    R. Panigrahy
KL Divergence Estimation with Multi-group Attribution
多组归因的 KL 散度估计
  • DOI:
  • 发表时间:
    2022
  • 期刊:
  • 影响因子:
    0
  • 作者:
    Parikshit Gopalan;Nina Narodytska;Omer Reingold;Vatsal Sharan;Udi Wieder
  • 通讯作者:
    Udi Wieder
Finding Heavily-Weighted Features in Data Streams
查找数据流中的重要特征
  • DOI:
  • 发表时间:
    2017
  • 期刊:
  • 影响因子:
    0
  • 作者:
    Kai Sheng Tai;Vatsal Sharan;Peter D. Bailis;G. Valiant
  • 通讯作者:
    G. Valiant

Vatsal Sharan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 }}

相似国自然基金

Computational Methods for Analyzing Toponome Data
  • 批准号:
    60601030
  • 批准年份:
    2006
  • 资助金额:
    17.0 万元
  • 项目类别:
    青年科学基金项目

相似海外基金

ProbAI: A Hub for the Mathematical and Computational Foundations of Probabilistic AI
ProbAI:概率人工智能的数学和计算基础中心
  • 批准号:
    EP/Y028783/1
  • 财政年份:
    2024
  • 资助金额:
    $ 55万
  • 项目类别:
    Research Grant
Collaborative Research: Foundations of programmable living materials through synthetic biofilm engineering and quantitative computational modeling
合作研究:通过合成生物膜工程和定量计算建模为可编程生物材料奠定基础
  • 批准号:
    2214021
  • 财政年份:
    2023
  • 资助金额:
    $ 55万
  • 项目类别:
    Standard Grant
CRCNS: Computational Foundations for Externalizing/Internalizing Psychopathology
CRCNS:外化/内化精神病理学的计算基础
  • 批准号:
    10831117
  • 财政年份:
    2023
  • 资助金额:
    $ 55万
  • 项目类别:
Collaborative Research: Foundations of programmable living materials through synthetic biofilm engineering and quantitative computational modeling
合作研究:通过合成生物膜工程和定量计算建模为可编程生物材料奠定基础
  • 批准号:
    2214020
  • 财政年份:
    2023
  • 资助金额:
    $ 55万
  • 项目类别:
    Standard Grant
Computational foundations of active visual sensing
主动视觉传感的计算基础
  • 批准号:
    10431247
  • 财政年份:
    2022
  • 资助金额:
    $ 55万
  • 项目类别:
Foundations of Computational Mathematics Conference – FoCM 2023
计算数学基础会议 – FoCM 2023
  • 批准号:
    2232812
  • 财政年份:
    2022
  • 资助金额:
    $ 55万
  • 项目类别:
    Standard Grant
Foundations of quantum computational advantage (FoQaCiA)
量子计算优势的基础 (FoQaCiA)
  • 批准号:
    569582-2021
  • 财政年份:
    2022
  • 资助金额:
    $ 55万
  • 项目类别:
    Alliance Grants
The foundations of quantum computational advantage
量子计算优势的基础
  • 批准号:
    RGPIN-2022-03103
  • 财政年份:
    2022
  • 资助金额:
    $ 55万
  • 项目类别:
    Discovery Grants Program - Individual
Computational Foundations of Machine Learning in the Era of Big Data
大数据时代机器学习的计算基础
  • 批准号:
    RGPIN-2017-05032
  • 财政年份:
    2022
  • 资助金额:
    $ 55万
  • 项目类别:
    Discovery Grants Program - Individual
The foundations of quantum computational advantage
量子计算优势的基础
  • 批准号:
    DGECR-2022-00354
  • 财政年份:
    2022
  • 资助金额:
    $ 55万
  • 项目类别:
    Discovery Launch Supplement
{{ showInfoDetail.title }}

作者:{{ showInfoDetail.author }}

知道了