CAREER: Penalty Logic for Structured Machine Learning
职业:结构化机器学习的惩罚逻辑
基本信息
- 批准号:0546867
- 负责人:
- 金额:$ 50万
- 依托单位:
- 依托单位国家:美国
- 项目类别:Continuing Grant
- 财政年份:2006
- 资助国家:美国
- 起止时间:2006-02-01 至 2012-01-31
- 项目状态:已结题
- 来源:
- 关键词:
项目摘要
Proposal 0546867"CAREER: Penalty Logic for Structured Machine Learning"PI: Alan FernOregon State UniversityThis research will study penalty logic as a knowledge representation technique for structured machine learning. Such learning problems involve inducing complex mappings between structured data types. Examples include learning to map American football video to play descriptions, and mapping the state of multi-agent planning problems to joint agent actions. Such problems often contain many "nearly sound" logical constraints, which are generally true, but sometimes violated. These constraints can be explicitly represented using penalty logic models, which are sets of weighted logical formulas, where each weight represents the cost of violating a formula. Penalty-logic models allow the synergistic combination of robust training methods for linear cost functions and years of work on logic-based representations. The project will study leveraging penalty logics in four directions: (1) learning model structure, (2) achieving practically efficient inference, (3) incorporating human provided knowledge, and (4) reducing labeling effort via active learning. The broader impact of this work will be to advance the applicability of structured machine learning to a wide range of interpretation and decision making problems, including those above. Planned educational activities include initiating an annual competition for Oregon high school students aimed at increasing CS enrollment and interest in AI.
建议0546867《职业生涯:结构化机器学习的惩罚逻辑》PI:Alan FernOregon州立大学本研究将研究惩罚逻辑作为结构化机器学习的知识表示技术。这样的学习问题涉及在结构化数据类型之间引入复杂的映射。例如,学习将美式足球视频映射到比赛描述,以及将多代理规划问题的状态映射到联合代理行动。这类问题通常包含许多“近乎合理”的逻辑约束,这些约束通常是正确的,但有时会被违反。这些约束可以使用惩罚逻辑模型显式表示,惩罚逻辑模型是加权逻辑公式集,其中每个权重表示违反公式的成本。惩罚逻辑模型允许将线性成本函数的稳健训练方法与多年基于逻辑的表示法的工作协同结合。该项目将从四个方向研究惩罚逻辑的杠杆作用:(1)学习模型结构,(2)实现实际有效的推理,(3)结合人类提供的知识,(4)通过主动学习减少标记工作量。这项工作的更广泛影响将是将结构化机器学习的适用性推进到包括上述问题在内的广泛的解释和决策问题中。计划的教育活动包括为俄勒冈州的高中生发起一年一度的竞赛,旨在增加CS入学人数和对人工智能的兴趣。
项目成果
期刊论文数量(0)
专著数量(0)
科研奖励数量(0)
会议论文数量(0)
专利数量(0)
数据更新时间:{{ journalArticles.updateTime }}
{{
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 }}
Alan Fern其他文献
Robust Learning for Adaptive Programs by Leveraging Program Structure
利用程序结构实现自适应程序的稳健学习
- DOI:
- 发表时间:
2010 - 期刊:
- 影响因子:0
- 作者:
Jervis Pinto;Alan Fern;Tim Bauer;Martin Erwig - 通讯作者:
Martin Erwig
Learning and transferring roles in multi-agent MDPs
多智能体 MDP 中的学习和角色转移
- DOI:
- 发表时间:
2008 - 期刊:
- 影响因子:0
- 作者:
Aaron Wilson;Alan Fern;Soumya Ray;Prasad Tadepalli - 通讯作者:
Prasad Tadepalli
The Origins of Common Sense in Humans and Machines
人类和机器常识的起源
- DOI:
- 发表时间:
2020 - 期刊:
- 影响因子:0
- 作者:
Kevin A. Smith;Eliza Kosoy;A. Gopnik;Deepak Pathak;Alan Fern;J. Tenenbaum;T. Ullman - 通讯作者:
T. Ullman
Active Imitation Learning via State Queries
通过状态查询进行主动模仿学习
- DOI:
- 发表时间:
2011 - 期刊:
- 影响因子:0
- 作者:
Kshitij Judah;Alan Fern - 通讯作者:
Alan Fern
Special report: The AgAID AI institute for transforming workforce and decision support in agriculture
- DOI:
10.1016/j.compag.2022.106944 - 发表时间:
2022-06-01 - 期刊:
- 影响因子:
- 作者:
Ananth Kalyanaraman;Margaret Burnett;Alan Fern;Lav Khot;Joshua Viers - 通讯作者:
Joshua Viers
Alan Fern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('Alan Fern', 18)}}的其他基金
Collaborative Research: CISE: Large: Executing Natural Instructions in Realistic Uncertain Worlds
合作研究:CISE:大型:在现实的不确定世界中执行自然指令
- 批准号:
2321851 - 财政年份:2023
- 资助金额:
$ 50万 - 项目类别:
Continuing Grant
Student Support for the 2020 International Conference on Automated Planning and Scheduling
2020 年自动规划与调度国际会议的学生支持
- 批准号:
2017913 - 财政年份:2020
- 资助金额:
$ 50万 - 项目类别:
Standard Grant
S&AS:INT:Learning and Planning for Dynamic Locomotion
S
- 批准号:
1849343 - 财政年份:2019
- 资助金额:
$ 50万 - 项目类别:
Standard Grant
RI: Small: Speedup Learning for Online Planning Under Uncertainty
RI:小:加速不确定性下在线规划的学习
- 批准号:
1619433 - 财政年份:2016
- 资助金额:
$ 50万 - 项目类别:
Standard Grant
II-EN: Software Tools for Monte-Carlo Optimization
II-EN:蒙特卡罗优化软件工具
- 批准号:
1406049 - 财政年份:2014
- 资助金额:
$ 50万 - 项目类别:
Standard Grant
RI: Small: Automated Planning of Experiments for Design Optimization
RI:小型:自动规划实验以优化设计
- 批准号:
1320943 - 财政年份:2013
- 资助金额:
$ 50万 - 项目类别:
Continuing Grant
Student Poster Program and Travel Scholarships for International Conference on Machine Learning (ICML) 2010; Haifa, Israel
2010 年国际机器学习会议 (ICML) 学生海报计划和旅行奖学金;
- 批准号:
1031917 - 财政年份:2010
- 资助金额:
$ 50万 - 项目类别:
Standard Grant
RI: Medium: Collaborative Research: Solving Stochastic Planning Problems Through Principled Determinization
RI:媒介:协作研究:通过原则确定解决随机规划问题
- 批准号:
0905678 - 财政年份:2009
- 资助金额:
$ 50万 - 项目类别:
Standard Grant
相似海外基金
The impact of Medicaid expansion on the rural mortality penalty in the United States
医疗补助扩大对美国农村死亡率的影响
- 批准号:
10726695 - 财政年份:2023
- 资助金额:
$ 50万 - 项目类别:
Promoting taxpayers' voluntary information disclosure under the self-assessment tax system and a new development of tax penalty and procedural legislations.
推动纳税人在自评税制下自愿信息披露以及税收处罚和程序立法的新发展。
- 批准号:
23K12370 - 财政年份:2023
- 资助金额:
$ 50万 - 项目类别:
Grant-in-Aid for Early-Career Scientists
Conservative discontinuous Galerkin methods with implicit penalty parameters and multiscale hybridizable discontinuous Galerkin methods for PDEs
具有隐式惩罚参数的保守间断伽辽金方法和偏微分方程的多尺度可杂交间断伽辽金方法
- 批准号:
2309670 - 财政年份:2023
- 资助金额:
$ 50万 - 项目类别:
Standard Grant
RAPID: Impact of Changing Death Penalty Unanimity Requirements on Case Decisions
RAPID:改变死刑一致要求对案件判决的影响
- 批准号:
2323006 - 财政年份:2023
- 资助金额:
$ 50万 - 项目类别:
Standard Grant
Penalty, non-penalty shrinkage and ensemble methods for low and high dimensional data
低维和高维数据的惩罚、非惩罚收缩和集成方法
- 批准号:
RGPIN-2019-04101 - 财政年份:2022
- 资助金额:
$ 50万 - 项目类别:
Discovery Grants Program - Individual
Mapping the Political Economy of Drugs and the Death Penalty in Southeast Asia
绘制东南亚毒品和死刑的政治经济图
- 批准号:
ES/W011425/1 - 财政年份:2022
- 资助金额:
$ 50万 - 项目类别:
Research Grant
Ensemble subspace, penalty, pretest, and shrinkage strategies for high dimensional data
高维数据的集成子空间、惩罚、预测试和收缩策略
- 批准号:
RGPIN-2017-05228 - 财政年份:2022
- 资助金额:
$ 50万 - 项目类别:
Discovery Grants Program - Individual
Climate Penalty: Climate-driven Increases in Ozone and PM2.5 Levels and Mortality
气候惩罚:气候驱动的臭氧和 PM2.5 水平和死亡率增加
- 批准号:
10218738 - 财政年份:2021
- 资助金额:
$ 50万 - 项目类别:
Climate Penalty: Climate-driven Increases in Ozone and PM2.5 Levels and Mortality
气候惩罚:气候驱动的臭氧和 PM2.5 水平和死亡率增加
- 批准号:
10372176 - 财政年份:2021
- 资助金额:
$ 50万 - 项目类别:
Penalty, non-penalty shrinkage and ensemble methods for low and high dimensional data
低维和高维数据的惩罚、非惩罚收缩和集成方法
- 批准号:
RGPIN-2019-04101 - 财政年份:2021
- 资助金额:
$ 50万 - 项目类别:
Discovery Grants Program - Individual