FAI: Towards Holistic Bias Mitigation in Computer Vision Systems
FAI: Towards Holistic Bias Mitigation in Computer Vision Systems
批准号:
2041009
负责人:
Nuno Vasconcelos
金额:
$37.5万
依托单位国家:
美国
项目类别:
Standard Grant
财政年份:
2021
资助国家:
美国
项目状态:
已结题
起止时间:
2021-02-01 至 2024-01-31
中文摘要
随着人工智能(AI)系统越来越多地用于改变生活的决策,如雇用或解雇个人或监禁刑期的长短,人们越来越关注这些系统的公平性。有必要保证人工智能系统不会对部分人口产生偏见。该项目旨在缓解计算机视觉领域的人工智能偏见,计算机视觉是一种热门的人工智能形式,即深度学习,这是一种推动最近许多进展的应用程序。计算机视觉系统在从医疗保健到执法的社会领域越来越普遍:从分析皮肤图片以检测黑色素瘤的应用程序,到用于刑事调查的人脸识别系统。这些系统有三个主要的偏差来源:有偏差的数据、有偏差的注释和有偏差的模型。有偏见的数据源于糟糕的图像收集做法,通常是某些人口群体的代表性不足。有偏见的注解源于使用带有未经训练的图像标签器的注解平台,这些注释器倾向于产生反映其自身图像解释的注解,而不是客观的标签。有偏差的模型可能源于对用于训练模型的数据集存在数据或注释偏差,或者选择有偏差的模型体系结构。这三个偏差分量在文献中受到了不同程度的关注,以往的大多数工作都集中在模型偏差的缓解上。然而,这通常归结为淡化有大量数据的小组,而促进数据稀缺的小组。这种做法可能会损害系统的整体性能。其余的偏倚来源,数据集和注释,在算法上几乎没有受到关注。该项目旨在克服这一问题,通过引入一个新的框架,在一个统一的偏差缓解架构内联合解决三个偏差来源。该体系结构旨在通过迭代优化三个不同的模块来训练公平的分类器:1)数据集偏差缓解算法,该算法识别和降低有偏见的样本,并在大量数据中寻找额外的样本来抵消相关的偏见。2)基于机器教学算法的标签偏差缓解系统,该系统建立了清晰、可复制和可审计的程序,以教注员如何在没有标签偏差的情况下对图像进行标签。3)基于反事实的可视解释的模型审计技术,能够可视化影响模型决策的因素以及它们为什么有偏差。这三个模块结合成一个架构,通过迭代优化数据集、注释器和模型以最大限度地减少偏差,从而缓解联合数据集、标签和模型的偏差。该项目将生成用于缓解数据集偏差、无偏见的注释员培训、解释和可视化、模型审计和公平模型培训的软件,这些软件将在调查员网站上提供。这将辅之以设计各种形式的偏差缓解算法的数据集,以及帮助从业者检测和打击偏差的工具。还计划开展几项活动,以扩大代表不足的K-12和本科生在STEM领域的参与。他们将包括从加州大学圣地亚哥分校项目中招募的此类学生团队的参与,该项目旨在增加这些群体在STEM中的参与,旨在为这些学生提供早期接触真实世界工程、公平机器学习和深度学习系统挑战的机会。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。
英文摘要
With the increasing use of artificial intelligence (AI) systems in life-changing decisions, such as hiring or firing of individuals or the length of jail sentences, there has been an increasing concern about the fairness of these systems. There is a need to guarantee that AI systems are not biased against segments of the population. This project aims to mitigate AI bias in the domain of computer vision, a driving application for much of the recent advances in a popular form of AI known as deep learning. Computer vision systems are increasingly prevalent in areas of society ranging from healthcare to law enforcement: from apps that analyze skin pictures for melanoma detection to face recognition systems used in criminal investigations. These systems are subject to three major sources of bias: biased data, biased annotations, and biased models. Biased data follows from poor image collection practices, typically the under-representation of certain population groups. Biased annotation follows from the use of annotation platforms with untrained image labelers, who tend to produce annotations that reflect their own image interpretations, rather than objective labels. Biased models can ensue from either the existence of data or annotation biases on the datasets used to train the models, or the choice of biased model architectures. The three bias components have received different attention in the literature, with most previous work focusing on the mitigation of model bias. However, this usually boils down to downplaying groups for which there is a lot of data and promoting groups for which data is scarce. This practice can hurt overall system performance. The remaining sources of bias, datasets and annotation, have received very little algorithmic attention. The project aims to overcome this problem, by introducing a new framework to jointly address the three sources of bias within one unified bias mitigation architecture. This architecture aims to train fair classifiers by iterative optimization of three distinct modules: 1) Dataset bias mitigation algorithms that identify and downweigh biased examples and seek additional examples in a large pool of data to counterbalance the associated biases. 2) Label bias mitigation systems based on machine teaching algorithms that establish clear, replicable, and auditable procedures to teach annotators how to label images without label bias. 3) Model auditing techniques based on counterfactual visual explanations that enable the visualization of the factors contributing to model decisions and why they are biased. The three modules combine into an architecture for joint dataset, label, and model bias mitigation by iterative optimization of datasets, annotators, and models to minimize bias. The project will generate software for dataset bias mitigation, unbiased annotator training, explanations and visualizations, model auditing, and fair model training, which will be made available from the investigator website. This will be complemented with datasets for the design of various form of bias mitigation algorithms, and tools to help practitioners detect and combat bias. Several activities are also planned to broaden the participation of underrepresented K-12 and undergraduate students in the STEM field. They will include the participation of a team of such students, recruited from University of California San Diego programs that aim to increase the participation of these groups in STEM, and aim to provide these students with early exposure to the challenges of real-world engineering, fair machine learning, and deep learning systems.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.
期刊论文(9)
专著(0)
科研奖励(0)
会议论文
登录
查看更多内容
DOI:
10.1109/cvpr52729.2023.00309
发表时间:
2023-06
期刊:
2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
影响因子:
--
作者:
[Pei Wang;N. Vasconcelos]
通讯作者:
Pei Wang;N. Vasconcelos
DOI:
10.1109/cvpr42600.2020.00900
发表时间:
2020-04
期刊:
2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
影响因子:
--
作者:
[Pei Wang;N. Vasconcelos]
通讯作者:
Pei Wang;N. Vasconcelos
DOI:
10.1109/cvpr52688.2022.00515
发表时间:
2022-05
期刊:
2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
影响因子:
--
作者:
[Yi Li-;Rameswar Panda;Yoon Kim;Chun-Fu Chen;R. Feris;David Cox;N. Vasconcelos]
通讯作者:
Yi Li-;Rameswar Panda;Yoon Kim;Chun-Fu Chen;R. Feris;David Cox;N. Vasconcelos
DOI:
--
发表时间:
2023
期刊:
IEEE/CVF International Conference on Computer Vision
影响因子:
--
作者:
[Ho, Chih-Hui, Zhang, Yuwei, Vasconcelos, Nuno]
通讯作者:
Vasconcelos, Nuno
Improving Video Model Transfer with Dynamic Representation Learning
通过动态表示学习改进视频模型传输
DOI:
--
发表时间:
2022
期刊:
IEEE/CVF Conference on Computer Vision and Pattern Recognition
影响因子:
--
作者:
[Yi Li, Nuno Vasconcelos]
通讯作者:
Yi Li, Nuno Vasconcelos
共 9 条
RI:Small:Dynamic Networks for Efficient, Adaptive, and Multimodal Vision
-
批准号:2303153
-
项目类别:Standard Grant
-
资助金额:$60.0万
-
财政年份:2023
-
负责人:Nuno Vasconcelos
-
依托单位:
NRI: FND: Towards Scalable and Self-Aware Robotic Perception
-
批准号:1924937
-
项目类别:Standard Grant
-
资助金额:$75.0万
-
财政年份:2019
-
负责人:Nuno Vasconcelos
-
依托单位:
NRI: Real-Time Semantic Computer Vision for Co-Robotics
-
批准号:1637941
-
项目类别:Standard Grant
-
资助金额:$71.91万
-
财政年份:2016
-
负责人:Nuno Vasconcelos
-
依托单位:
BIGDATA: Collaborative Research: IA: Quantifying Plankton Diversity with Taxonomy and Attribute Based Classifiers of Underwater Microscope Images
-
批准号:1546305
-
项目类别:Standard Grant
-
资助金额:$28.34万
-
财政年份:2016
-
负责人:Nuno Vasconcelos
-
依托单位:
NRI-Small: A Biologically Plausible Architecture for Robotic Vision
-
批准号:1208522
-
项目类别:Standard Grant
-
资助金额:$115.0万
-
财政年份:2012
-
负责人:Nuno Vasconcelos
-
依托单位:
Large-vocabulary Semantic Image Processing: Theory and Algorithms
-
批准号:0830535
-
项目类别:Standard Grant
-
资助金额:$23.95万
-
财政年份:2008
-
负责人:Nuno Vasconcelos
-
依托单位:
RI-Small: Optimal Automated Design of Cascaded Object Detectors
-
批准号:0812235
-
项目类别:Standard Grant
-
资助金额:$33.7万
-
财政年份:2008
-
负责人:Nuno Vasconcelos
-
依托单位:
Understanding Video of Crowded Environments
-
批准号:0534985
-
项目类别:Continuing Grant
-
资助金额:$0.0万
-
财政年份:2005
-
负责人:Nuno Vasconcelos
-
依托单位:
CAREER: Weakly Supervised Recognition
-
批准号:0448609
-
项目类别:Continuing Grant
-
资助金额:$0.0万
-
财政年份:2005
-
负责人:Nuno Vasconcelos
-
依托单位:
海外基金