近年来,关于如何在公平限制下学习机器学习模型的越来越多的工作,通常在某些敏感属性方面表达。在这项工作中,我们考虑了对手对目标模型具有黑箱访问的设置,并表明对手可以利用有关该模型公平性的信息,以增强他对训练数据敏感属性的重建。更确切地说,我们提出了一种通用的重建校正方法,该方法将其作为对手进行的初始猜测,并纠正它以符合某些用户定义的约束(例如公平信息),同时最大程度地减少了对手猜测的变化。提出的方法对目标模型的类型,公平感知的学习方法以及对手的辅助知识不可知。为了评估我们的方法的适用性,我们对两种最先进的公平学习方法进行了彻底的实验评估,使用四个具有广泛公差的不同公平指标以及三个不同大小和敏感属性的数据集。实验结果证明了提出的方法改善训练集敏感属性的重建的有效性。
translated by 谷歌翻译
为了减轻模型中不希望的偏差的影响,几种方法建议预先处理输入数据集,以通过防止敏感属性的推断来减少歧视风险。不幸的是,这些预处理方法中的大多数导致一代新分布与原始分布有很大不同,因此通常导致不切实际的数据。作为副作用,这种新的数据分布意味着需要重新训练现有模型才能做出准确的预测。为了解决这个问题,我们提出了一种新颖的预处理方法,我们将根据保护组的分布转换为所选目标一个,并具有附加的隐私约束,其目的是防止敏感敏感的推断属性。更确切地说,我们利用Wasserstein Gan和Attgan框架的最新作品来实现数据点的最佳运输以及强制保护属性推断的歧视器。我们提出的方法可以保留数据的可解释性,并且可以在不定义敏感组的情况下使用。此外,我们的方法可以专门建模现有的最新方法,从而提出对这些方法的统一观点。最后,关于真实和合成数据集的一些实验表明,我们的方法能够隐藏敏感属性,同时限制数据的变形并改善了后续数据分析任务的公平性。
translated by 谷歌翻译
分类,一种重大研究的数据驱动机器学习任务,驱动越来越多的预测系统,涉及批准的人类决策,如贷款批准和犯罪风险评估。然而,分类器经常展示歧视性行为,特别是当呈现有偏置数据时。因此,分类公平已经成为一个高优先级的研究区。数据管理研究显示与数据和算法公平有关的主题的增加和兴趣,包括公平分类的主题。公平分类的跨学科努力,具有最大存在的机器学习研究,导致大量的公平概念和尚未系统地评估和比较的广泛方法。在本文中,我们对13个公平分类方法和额外变种的广泛分析,超越,公平,公平,效率,可扩展性,对数据误差的鲁棒性,对潜在的ML模型,数据效率和使用各种指标的稳定性的敏感性和稳定性现实世界数据集。我们的分析突出了对不同指标的影响的新颖见解和高级方法特征对不同方面的性能方面。我们还讨论了选择适合不同实际设置的方法的一般原则,并确定以数据管理为中心的解决方案可能产生最大影响的区域。
translated by 谷歌翻译
本文考虑了在分解正常形式(DNF,ANDS的DNF,ANDS,相当于判定规则集)或联合正常形式(CNF,ORS)作为分类模型的联合正常形式的学习。为规则简化,将整数程序配制成最佳贸易分类准确性。我们还考虑公平设定,并扩大制定,以包括对两种不同分类措施的明确限制:机会平等和均等的赔率。列生成(CG)用于有效地搜索候选条款(连词或剖钉)的指数数量,而不需要启发式规则挖掘。此方法还会绑定所选规则集之间的间隙和培训数据上的最佳规则集。要处理大型数据集,我们建议使用随机化的近似CG算法。与三个最近提出的替代方案相比,CG算法主导了16个数据集中的8个中的精度简单折衷。当最大限度地提高精度时,CG与为此目的设计的规则学习者具有竞争力,有时发现明显更简单的解决方案,这些解决方案不太准确。与其他公平和可解释的分类器相比,我们的方法能够找到符合较严格的公平概念的规则集,以适度的折衷准确性。
translated by 谷歌翻译
机器学习模型在许多领域都表现出了有希望的表现。但是,担心他们可能会偏向特定的群体,阻碍了他们在高级申请中的采用。因此,必须确保机器学习模型中的公平性。以前的大多数努力都需要访问敏感属性以减轻偏见。尽管如此,由于人们对隐私和法律依从性的认识日益增加,获得具有敏感属性的大规模数据通常是不可行的。因此,一个重要的研究问题是如何在隐私下做出公平的预测?在本文中,我们研究了半私人环境中公平分类的新问题,其中大多数敏感属性都是私有的,只有少量的干净敏感属性可用。为此,我们提出了一个新颖的框架Fairsp,可以首先学会通过利用有限的清洁敏感属性来纠正隐私保证下的嘈杂敏感属性。然后,它以对抗性方式共同建模校正和清洁数据以进行歧义和预测。理论分析表明,当大多数敏感属性都是私有的时,提出的模型可以确保公平。现实世界数据集的实验结果证明了所提出的模型在隐私下做出公平预测并保持高精度的有效性。
translated by 谷歌翻译
在高风险领域(人们的生计受到影响)中,机器学习的日益增长的使用迫切需要解释和公平的算法。在这些设置中,此类算法的准确性也至关重要。考虑到这些需求,我们提出了一个混合整数优化(MIO)框架,用于学习具有固定深度的最佳分类树,可以通过任意域特定的公平约束来方便地增强。我们基于在流行数据集上建造公平树木的最先进方法基准测试;鉴于固定的歧视阈值,我们的方法平均将样本外(OOS)的精度提高了2.3个百分点,并在88.9%的实验上获得了更高的OOS精度。我们还将各种算法公平概念纳入我们的方法中,展示其多功能建模能力,使决策者可以微调准确性和公平性之间的权衡。
translated by 谷歌翻译
机器学习模型在高赌注应用中变得普遍存在。尽管在绩效方面有明显的效益,但该模型可以表现出对少数民族群体的偏见,并导致决策过程中的公平问题,导致对个人和社会的严重负面影响。近年来,已经开发了各种技术来减轻机器学习模型的偏差。其中,加工方法已经增加了社区的关注,在模型设计期间直接考虑公平,以诱导本质上公平的模型,从根本上减轻了产出和陈述中的公平问题。在本调查中,我们审查了加工偏置减缓技术的当前进展。基于在模型中实现公平的地方,我们将它们分类为明确和隐性的方法,前者直接在培训目标中纳入公平度量,后者重点介绍精炼潜在代表学习。最后,我们在讨论该社区中的研究挑战来讨论调查,以激励未来的探索。
translated by 谷歌翻译
鉴于对机器学习模型的访问,可以进行对手重建模型的培训数据?这项工作从一个强大的知情对手的镜头研究了这个问题,他们知道除了一个之外的所有培训数据点。通过实例化混凝土攻击,我们表明重建此严格威胁模型中的剩余数据点是可行的。对于凸模型(例如Logistic回归),重建攻击很简单,可以以封闭形式导出。对于更常规的模型(例如神经网络),我们提出了一种基于训练的攻击策略,该攻击策略接收作为输入攻击的模型的权重,并产生目标数据点。我们展示了我们对MNIST和CIFAR-10训练的图像分类器的攻击的有效性,并系统地研究了标准机器学习管道的哪些因素影响重建成功。最后,我们从理论上调查了有多差异的隐私足以通过知情对手减轻重建攻击。我们的工作提供了有效的重建攻击,模型开发人员可以用于评估超出以前作品中考虑的一般设置中的个别点的记忆(例如,生成语言模型或访问培训梯度);它表明,标准模型具有存储足够信息的能力,以实现培训数据点的高保真重建;它表明,差异隐私可以成功减轻该参数制度中的攻击,其中公用事业劣化最小。
translated by 谷歌翻译
在培训机器学习模型期间,它们可能会存储或“了解”有关培训数据的更多信息,而不是预测或分类任务所需的信息。属性推理攻击旨在从给定模型的培训数据中提取统计属性,而无需访问培训数据本身,从而利用了这一点。这些属性可能包括图片的质量,以识别相机模型,以揭示产品的目标受众的年龄分布或在计算机网络中使用恶意软件攻击的随附的主机类型。当攻击者可以访问所有模型参数时,即在白色盒子方案中,此攻击尤其准确。通过捍卫此类攻击,模型所有者可以确保其培训数据,相关的属性以及其知识产权保持私密,即使他们故意共享自己的模型,例如协作培训或模型泄漏。在本文中,我们介绍了属性,这是针对白盒属性推理攻击的有效防御机制,独立于培训数据类型,模型任务或属性数量。属性通过系统地更改目标模型的训练的权重和偏见来减轻属性推理攻击,从而使对手无法提取所选属性。我们在三个不同的数据集(包括表格数据和图像数据)以及两种类型的人工神经网络(包括人造神经网络)上进行了经验评估属性。我们的研究结果表明,以良好的隐私性权衡取舍,可以保护机器学习模型免受财产推理攻击的侵害,既有效又可靠。此外,我们的方法表明该机制也有效地取消了多个特性。
translated by 谷歌翻译
At the core of insurance business lies classification between risky and non-risky insureds, actuarial fairness meaning that risky insureds should contribute more and pay a higher premium than non-risky or less-risky ones. Actuaries, therefore, use econometric or machine learning techniques to classify, but the distinction between a fair actuarial classification and "discrimination" is subtle. For this reason, there is a growing interest about fairness and discrimination in the actuarial community Lindholm, Richman, Tsanakas, and Wuthrich (2022). Presumably, non-sensitive characteristics can serve as substitutes or proxies for protected attributes. For example, the color and model of a car, combined with the driver's occupation, may lead to an undesirable gender bias in the prediction of car insurance prices. Surprisingly, we will show that debiasing the predictor alone may be insufficient to maintain adequate accuracy (1). Indeed, the traditional pricing model is currently built in a two-stage structure that considers many potentially biased components such as car or geographic risks. We will show that this traditional structure has significant limitations in achieving fairness. For this reason, we have developed a novel pricing model approach. Recently some approaches have Blier-Wong, Cossette, Lamontagne, and Marceau (2021); Wuthrich and Merz (2021) shown the value of autoencoders in pricing. In this paper, we will show that (2) this can be generalized to multiple pricing factors (geographic, car type), (3) it perfectly adapted for a fairness context (since it allows to debias the set of pricing components): We extend this main idea to a general framework in which a single whole pricing model is trained by generating the geographic and car pricing components needed to predict the pure premium while mitigating the unwanted bias according to the desired metric.
translated by 谷歌翻译
A distribution inference attack aims to infer statistical properties of data used to train machine learning models. These attacks are sometimes surprisingly potent, but the factors that impact distribution inference risk are not well understood and demonstrated attacks often rely on strong and unrealistic assumptions such as full knowledge of training environments even in supposedly black-box threat scenarios. To improve understanding of distribution inference risks, we develop a new black-box attack that even outperforms the best known white-box attack in most settings. Using this new attack, we evaluate distribution inference risk while relaxing a variety of assumptions about the adversary's knowledge under black-box access, like known model architectures and label-only access. Finally, we evaluate the effectiveness of previously proposed defenses and introduce new defenses. We find that although noise-based defenses appear to be ineffective, a simple re-sampling defense can be highly effective. Code is available at https://github.com/iamgroot42/dissecting_distribution_inference
translated by 谷歌翻译
从公共机器学习(ML)模型中泄漏数据是一个越来越重要的领域,因为ML的商业和政府应用可以利用多个数据源,可能包括用户和客户的敏感数据。我们对几个方面的当代进步进行了全面的调查,涵盖了非自愿数据泄漏,这对ML模型很自然,潜在的恶毒泄漏是由隐私攻击引起的,以及目前可用的防御机制。我们专注于推理时间泄漏,这是公开可用模型的最可能场景。我们首先在不同的数据,任务和模型体系结构的背景下讨论什么是泄漏。然后,我们提出了跨非自愿和恶意泄漏的分类法,可用的防御措施,然后进行当前可用的评估指标和应用。我们以杰出的挑战和开放性的问题结束,概述了一些有希望的未来研究方向。
translated by 谷歌翻译
住院患者的高血糖治疗对发病率和死亡率都有重大影响。这项研究使用了大型临床数据库来预测需要住院的糖尿病患者的需求,这可能会改善患者的安全性。但是,这些预测可能容易受到社会决定因素(例如种族,年龄和性别)造成的健康差异的影响。这些偏见必须在数据收集过程的早期,在进入系统之前就可以消除,并通过模型预测加强,从而导致模型决策的偏见。在本文中,我们提出了一条能够做出预测以及检测和减轻偏见的机器学习管道。该管道分析了临床数据,确定是否存在偏见,将其删除,然后做出预测。我们使用实验证明了模型预测中的分类准确性和公平性。结果表明,当我们在模型早期减轻偏见时,我们会得到更公平的预测。我们还发现,随着我们获得更好的公平性,我们牺牲了一定程度的准确性,这在先前的研究中也得到了验证。我们邀请研究界为确定可以通过本管道解决的其他因素做出贡献。
translated by 谷歌翻译
模型可以公开有关其培训数据的敏感信息。在属性推理攻击中,对手对某些培训记录有部分知识,并访问了对这些记录进行培训的模型,并渗透了这些记录敏感功能的未知值。我们研究了一种属性推理的细粒变体,我们称为\ emph {敏感值推理},其中对手的目标是高度置信度识别一些来自候选人集的记录,其中未知属性具有特定的敏感值。我们将属性推断与捕获培训分布统计数据的数据插补进行明确比较,该数据在对对手可用的培训数据的各种假设下进行了比较。我们的主要结论是:(1)以前的属性推理方法并没有比对手可以推断出有关训练数据的训练数据的更多信息,而无需访问训练的模型,而是对培训所需的基础分布相同的知识属性推理攻击; (2)Black-Box属性推理攻击很少学习没有模型的任何东西;但是(3)我们在论文中介绍和评估的白框攻击可以可靠地识别一些具有敏感值属性的记录,而这些记录在不访问模型的情况下无法预测。此外,我们表明提出的防御措施,例如私人培训和从培训中删除脆弱记录不会减轻这种隐私风险。我们的实验代码可在\ url {https://github.com/bargavj/evaluatingdpml}上获得。
translated by 谷歌翻译
A recent explosion of research focuses on developing methods and tools for building fair predictive models. However, most of this work relies on the assumption that the training and testing data are representative of the target population on which the model will be deployed. However, real-world training data often suffer from selection bias and are not representative of the target population for many reasons, including the cost and feasibility of collecting and labeling data, historical discrimination, and individual biases. In this paper, we introduce a new framework for certifying and ensuring the fairness of predictive models trained on biased data. We take inspiration from query answering over incomplete and inconsistent databases to present and formalize the problem of consistent range approximation (CRA) of answers to queries about aggregate information for the target population. We aim to leverage background knowledge about the data collection process, biased data, and limited or no auxiliary data sources to compute a range of answers for aggregate queries over the target population that are consistent with available information. We then develop methods that use CRA of such aggregate queries to build predictive models that are certifiably fair on the target population even when no external information about that population is available during training. We evaluate our methods on real data and demonstrate improvements over state of the art. Significantly, we show that enforcing fairness using our methods can lead to predictive models that are not only fair, but more accurate on the target population.
translated by 谷歌翻译
随着机器学习变得普遍,减轻培训数据中存在的任何不公平性变得至关重要。在公平的各种概念中,本文的重点是众所周知的个人公平,该公平规定应该对类似的人进行类似的对待。虽然在训练模型(对处理)时可以提高个人公平性,但我们认为在模型培训(预处理)之前修复数据是一个更基本的解决方案。特别是,我们表明标签翻转是改善个人公平性的有效预处理技术。我们的系统IFLIPPER解决了限制了个人公平性违规行为的最小翻转标签的优化问题,当培训数据中的两个类似示例具有不同的标签时,发生违规情况。我们首先证明问题是NP-HARD。然后,我们提出了一种近似的线性编程算法,并提供理论保证其结果与标签翻转数量有关的结果与最佳解决方案有多近。我们还提出了使线性编程解决方案更加最佳的技术,而不会超过违规限制。实际数据集上的实验表明,在看不见的测试集的个人公平和准确性方面,IFLIPPER显着优于其他预处理基线。此外,IFLIPPER可以与处理中的技术结合使用,以获得更好的结果。
translated by 谷歌翻译
我们解决了分类中群体公平的问题,目的是学习不会不公正地歧视人口亚组的模型。大多数现有方法仅限于简单的二进制任务或涉及难以实施培训机制。这降低了他们的实际适用性。在本文中,我们提出了Fairgrad,这是一种基于重新加权方案来实施公平性的方法,该计划根据是否有优势地迭代地学习特定权重。Fairgrad易于实施,可以适应各种标准公平定义。此外,我们表明它与各种数据集的标准基线相媲美,包括自然语言处理和计算机视觉中使用的数据集。
translated by 谷歌翻译
Existing regulations prohibit model developers from accessing protected attributes (gender, race, etc.), often resulting in fairness assessments on populations without knowing their protected groups. In such scenarios, institutions often adopt a separation between the model developers (who train models with no access to the protected attributes) and a compliance team (who may have access to the entire dataset for auditing purpose). However, the model developers might be allowed to test their models for bias by querying the compliance team for group fairness metrics. In this paper, we first demonstrate that simply querying for fairness metrics, such as statistical parity and equalized odds can leak the protected attributes of individuals to the model developers. We demonstrate that there always exist strategies by which the model developers can identify the protected attribute of a targeted individual in the test dataset from just a single query. In particular, we show that one can reconstruct the protected attributes of all the individuals from O(Nk log n/Nk) queries when Nk<<n using techniques from compressed sensing (n: size of the test dataset, Nk: size of smallest group). Our results pose an interesting debate in algorithmic fairness: should querying for fairness metrics be viewed as a neutral-valued solution to ensure compliance with regulations? Or, does it constitute a violation of regulations and privacy if the number of queries answered is enough for the model developers to identify the protected attributes of specific individuals? To address this supposed violation, we also propose Attribute-Conceal, a novel technique that achieves differential privacy by calibrating noise to the smooth sensitivity of our bias query, outperforming naive techniques such as Laplace mechanism. We also include experimental results on the Adult dataset and synthetic data (broad range of parameters).
translated by 谷歌翻译
Although query-based systems (QBS) have become one of the main solutions to share data anonymously, building QBSes that robustly protect the privacy of individuals contributing to the dataset is a hard problem. Theoretical solutions relying on differential privacy guarantees are difficult to implement correctly with reasonable accuracy, while ad-hoc solutions might contain unknown vulnerabilities. Evaluating the privacy provided by QBSes must thus be done by evaluating the accuracy of a wide range of privacy attacks. However, existing attacks require time and expertise to develop, need to be manually tailored to the specific systems attacked, and are limited in scope. In this paper, we develop QuerySnout (QS), the first method to automatically discover vulnerabilities in QBSes. QS takes as input a target record and the QBS as a black box, analyzes its behavior on one or more datasets, and outputs a multiset of queries together with a rule to combine answers to them in order to reveal the sensitive attribute of the target record. QS uses evolutionary search techniques based on a novel mutation operator to find a multiset of queries susceptible to lead to an attack, and a machine learning classifier to infer the sensitive attribute from answers to the queries selected. We showcase the versatility of QS by applying it to two attack scenarios, three real-world datasets, and a variety of protection mechanisms. We show the attacks found by QS to consistently equate or outperform, sometimes by a large margin, the best attacks from the literature. We finally show how QS can be extended to QBSes that require a budget, and apply QS to a simple QBS based on the Laplace mechanism. Taken together, our results show how powerful and accurate attacks against QBSes can already be found by an automated system, allowing for highly complex QBSes to be automatically tested "at the pressing of a button".
translated by 谷歌翻译
We quantitatively investigate how machine learning models leak information about the individual data records on which they were trained. We focus on the basic membership inference attack: given a data record and black-box access to a model, determine if the record was in the model's training dataset. To perform membership inference against a target model, we make adversarial use of machine learning and train our own inference model to recognize differences in the target model's predictions on the inputs that it trained on versus the inputs that it did not train on.We empirically evaluate our inference techniques on classification models trained by commercial "machine learning as a service" providers such as Google and Amazon. Using realistic datasets and classification tasks, including a hospital discharge dataset whose membership is sensitive from the privacy perspective, we show that these models can be vulnerable to membership inference attacks. We then investigate the factors that influence this leakage and evaluate mitigation strategies.
translated by 谷歌翻译