CAREER: User-Based Simulation Methods for Quantifying Sources of Error and Bias in Recommender Systems
职业:基于用户的模拟方法,用于量化推荐系统中的错误和偏差来源
基本信息
- 批准号:2415042
- 负责人:
- 金额:$ 48.21万
- 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:2023-10-01 至 2024-08-31
- 项目状态:已结题
- 来源:
- 关键词:
项目摘要
Systems that recommend products, places, and services are an increasingly common part of everyday life and commerce, making it important to understand how recommendation algorithms affect outcomes for both individual users and larger social groups. To do this, the project team will develop novel methods of simulating users' behavior based on large-scale historical datasets. These methods will be used to better understand vulnerabilities that underlying biases in training datasets pose to commonly-used machine learning-based methods for building and testing recommender systems, as well as characterize the effectiveness of common evaluation metrics such as recommendation accuracy and diversity given different models of how people interact with recommender systems in practice. The team will publicly release its datasets, software, and novel metrics for the benefit of other researchers and developers of recommender systems. The work also will inform the development of computer science course materials about the social impact of data analytics as well as outreach activities for librarians, who are often in the position of helping information seekers understand the way search engines and other recommender systems affect their ability to get what they need.The work is organized around two main themes. The first will quantify and mitigate the popularity bias and misclassified decoy problems in offline recommender evaluation that tend to lead to popular, known recommendations. To do this, the team will develop simulation-based evaluation models that encode a variety of assumptions about how users select relevant items to buy and rate and use them to quantify the statistical biases these assumptions induce in recommendation quality metrics. They will calibrate these simulations by comparing with existing data sets covering books, research papers, music, and movies. These models and datasets will help drive the second main project around measuring the impact of feature distributions in training data on recommender algorithm accuracy and diversity, while developing bias-resistant algorithms. The team will use data resampling techniques along with the simulation models, extended to model system behavior over time, to evaluate how different algorithms mitigate, propagate, or exacerbate underlying distributional biases through their recommendations, and how those biased recommendations in turn affect future user behavior and experience.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.
推荐产品、地点和服务的系统在日常生活和商业中越来越常见,因此了解推荐算法如何影响个人用户和更大的社会群体的结果变得非常重要。 为此,项目团队将开发基于大规模历史数据集模拟用户行为的新方法。 这些方法将用于更好地理解训练数据集中的潜在偏见对构建和测试推荐系统的常用机器学习方法造成的脆弱性,以及描述常见评估指标的有效性,例如推荐准确性和多样性,给出人们在实践中如何与推荐系统交互的不同模型。 该团队将公开发布其数据集,软件和新的指标,以使其他研究人员和推荐系统开发人员受益。 这项工作还将为计算机科学课程材料的开发提供信息,这些材料涉及数据分析的社会影响以及图书馆员的外联活动,图书馆员通常处于帮助信息寻求者了解搜索引擎和其他推荐系统如何影响他们获得所需信息的能力的位置。 第一个将量化和减轻离线推荐评估中的流行偏见和错误分类诱饵问题,这些问题往往会导致流行的,已知的推荐。 为此,该团队将开发基于模拟的评估模型,该模型编码了关于用户如何选择购买和评级的相关项目的各种假设,并使用它们来量化这些假设在推荐质量指标中引起的统计偏差。他们将通过与现有的书籍、研究论文、音乐和电影数据集进行比较来校准这些模拟。 这些模型和数据集将有助于推动第二个主要项目,即测量训练数据中的特征分布对推荐算法准确性和多样性的影响,同时开发抗偏见算法。 该团队将使用数据恢复技术沿着模拟模型,扩展到随着时间的推移对系统行为进行建模,以评估不同的算法如何通过其建议减轻,传播或加剧潜在的分布偏差,以及这些有偏见的建议如何反过来影响未来的用户行为和体验。该奖项反映了NSF的法定使命,并被认为值得通过使用基金会的学术价值和更广泛的影响审查标准。
项目成果
期刊论文数量(3)
专著数量(0)
科研奖励数量(0)
会议论文数量(0)
专利数量(0)
Not Just Algorithms: Strategically Addressing Consumer Impacts in Information Retrieval
- DOI:10.1007/978-3-031-56066-8_25
- 发表时间:2024
- 期刊:
- 影响因子:0
- 作者:Michael D. Ekstrand;Lex Beattie;M. S. Pera;Henriette Cramer
- 通讯作者:Michael D. Ekstrand;Lex Beattie;M. S. Pera;Henriette Cramer
Multiple Testing for IR and Recommendation System Experiments
IR 和推荐系统实验的多重测试
- DOI:
- 发表时间:2024
- 期刊:
- 影响因子:0
- 作者:Ihemelandu, Ngozi;Ekstrand, Michael D.
- 通讯作者:Ekstrand, Michael D.
Towards Optimizing Ranking in Grid-Layout for Provider-side Fairness
优化网格布局排名以实现提供商方公平
- DOI:
- 发表时间:2024
- 期刊:
- 影响因子:0
- 作者:Raj, Amifa;Ekstrand, Michael D.
- 通讯作者:Ekstrand, Michael D.
{{
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 }}
Michael Ekstrand其他文献
Making Algorithms Public: Reimagining Auditing from Matters of Fact to Matters of Concern
公开算法:重新构想从事实问题到关注问题的审计
- DOI:
- 发表时间:
- 期刊:
- 影响因子:0
- 作者:
R. Stuart;Irani;L. Irani;Kathryne Metcalf;Magdalena Donea;Julia Kott;Jennifer Chien;Emma Jablonski;Christian Sandvig;Karrie Karrahalios;Peaks Krafft Meg Young;Mike Katell;Michael Ekstrand;Katie Shilton;C. Kelty;Kristen Vaccaro;Mary Anne - 通讯作者:
Mary Anne
Responsible AI Research Needs Impact Statements Too
负责任的人工智能研究也需要影响报告
- DOI:
- 发表时间:
2023 - 期刊:
- 影响因子:0
- 作者:
Alexandra Olteanu;Michael Ekstrand;Carlos Castillo;Jina Suh - 通讯作者:
Jina Suh
Michael Ekstrand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('Michael Ekstrand', 18)}}的其他基金
Collaborative Research: CCRI: New: A Research News Recommender Infrastructure with Live Users for Algorithm and Interface Experimentation
合作研究:CCRI:新:研究新闻推荐基础设施与实时用户进行算法和界面实验
- 批准号:
2232553 - 财政年份:2023
- 资助金额:
$ 48.21万 - 项目类别:
Standard Grant
Collaborative Research: CCRI: New: A Research News Recommender Infrastructure with Live Users for Algorithm and Interface Experimentation
合作研究:CCRI:新:研究新闻推荐基础设施与实时用户进行算法和界面实验
- 批准号:
2409199 - 财政年份:2023
- 资助金额:
$ 48.21万 - 项目类别:
Standard Grant
CAREER: User-Based Simulation Methods for Quantifying Sources of Error and Bias in Recommender Systems
职业:基于用户的模拟方法,用于量化推荐系统中的错误和偏差来源
- 批准号:
1751278 - 财政年份:2018
- 资助金额:
$ 48.21万 - 项目类别:
Continuing Grant
相似海外基金
The Lowdown - An AI-based medical device that uniquely combines clinical guidance with real user experience to support improved contraceptive understanding and use.
The Lowdown - 一种基于人工智能的医疗设备,将临床指导与真实用户体验独特地结合起来,以支持提高避孕药具的理解和使用。
- 批准号:
10103064 - 财政年份:2024
- 资助金额:
$ 48.21万 - 项目类别:
Collaborative R&D
Road Surface Evaluation based on Physiopsychological User Responses Corresponding to Diversified Mobilities on Pedestrian Spaces
基于行人空间多样化移动的用户生理心理反应的路面评价
- 批准号:
23K04056 - 财政年份:2023
- 资助金额:
$ 48.21万 - 项目类别:
Grant-in-Aid for Scientific Research (C)
Wheelchair user physical activity training intervention to enhance cardiometabolic health (WATCH): A community-based randomized control trial
轮椅使用者身体活动训练干预以增强心脏代谢健康(WATCH):一项基于社区的随机对照试验
- 批准号:
10598219 - 财政年份:2023
- 资助金额:
$ 48.21万 - 项目类别:
Designing Infectious Disease Data Submission Tool Based on User Feedback
根据用户反馈设计传染病数据提交工具
- 批准号:
486994 - 财政年份:2023
- 资助金额:
$ 48.21万 - 项目类别:
Miscellaneous Programs
Development and delivery of an evidence-based user-centred nutrition education program - Nutrition in Medical Education (NutriMed) for healthcare professionals to improve their confidence in providing nutritional guidance
为医疗保健专业人员开发和提供以用户为中心的循证营养教育计划 - 医学教育中的营养 (NutriMed),以提高他们提供营养指导的信心
- 批准号:
484625 - 财政年份:2023
- 资助金额:
$ 48.21万 - 项目类别:
Fellowship Programs
User-Centered Design of a Proactive RF-Based Wearable Bladder Monitor for Toilet Training of Children with ASD/IDD
以用户为中心的主动式射频可穿戴膀胱监测器设计,用于 ASD/IDD 儿童如厕训练
- 批准号:
10742670 - 财政年份:2023
- 资助金额:
$ 48.21万 - 项目类别:
Facilitating evidence-based decision-making to combat antimicrobial resistance: development of a public, user-friendly database of the highest levels of evidence
促进基于证据的决策以对抗抗菌素耐药性:开发一个公共的、用户友好的最高水平证据数据库
- 批准号:
494278 - 财政年份:2023
- 资助金额:
$ 48.21万 - 项目类别:
Operating Grants
Extending ezBIDS, NiiVue and dcm2niix for user-friendly cloud-based integration and visualization
扩展 ezBIDS、NiiVue 和 dcm2niix,以实现用户友好的基于云的集成和可视化
- 批准号:
10724895 - 财政年份:2023
- 资助金额:
$ 48.21万 - 项目类别:
Developing an Innovative Platform for Modeling Active Road User Interactions and Safety: Integration of Computer Vision, Agent-based, and Machine Learning Models
开发用于对主动道路用户交互和安全进行建模的创新平台:计算机视觉、基于代理和机器学习模型的集成
- 批准号:
RGPIN-2019-06688 - 财政年份:2022
- 资助金额:
$ 48.21万 - 项目类别:
Discovery Grants Program - Individual
CNS Core: Medium: A Systems and User-based Approach to Floating Point Correctness and Resilience
CNS 核心:中:基于系统和用户的浮点正确性和弹性方法
- 批准号:
2211315 - 财政年份:2022
- 资助金额:
$ 48.21万 - 项目类别:
Continuing Grant