It has been observed that large-scale language models capture undesirable societal biases, e.g. relating to race and gender; yet religious bias has been relatively unexplored. We demonstrate that GPT-3, a state-of-the-art contextual language model, captures persistent Muslim-violence bias. We probe GPT-3 in various ways, including prompt completion, analogical reasoning, and story generation, to understand this anti-Muslim bias, demonstrating that it appears consistently and creatively in different uses of the model and that it is severe even compared to biases about other religious groups. For instance, Muslim is analogized to terrorist in 23% of test cases, while Jewish is mapped to its most common stereotype, money, in 5% of test cases. We quantify the positive distraction needed to overcome this bias with adversarial text prompts, and find that use of the most positive 6 adjectives reduces violent completions for Muslims from 66% to 20%, but which is still higher than for other religious groups.
{"title":"Persistent Anti-Muslim Bias in Large Language Models","authors":"Abubakar Abid, Maheen Farooqi, James Y. Zou","doi":"10.1145/3461702.3462624","DOIUrl":"https://doi.org/10.1145/3461702.3462624","url":null,"abstract":"It has been observed that large-scale language models capture undesirable societal biases, e.g. relating to race and gender; yet religious bias has been relatively unexplored. We demonstrate that GPT-3, a state-of-the-art contextual language model, captures persistent Muslim-violence bias. We probe GPT-3 in various ways, including prompt completion, analogical reasoning, and story generation, to understand this anti-Muslim bias, demonstrating that it appears consistently and creatively in different uses of the model and that it is severe even compared to biases about other religious groups. For instance, Muslim is analogized to terrorist in 23% of test cases, while Jewish is mapped to its most common stereotype, money, in 5% of test cases. We quantify the positive distraction needed to overcome this bias with adversarial text prompts, and find that use of the most positive 6 adjectives reduces violent completions for Muslims from 66% to 20%, but which is still higher than for other religious groups.","PeriodicalId":197336,"journal":{"name":"Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130306437","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
We study whether receiving advice from either a human or algorithmic advisor, accompanied by five types of Local and Global explanation labelings, has an effect on the readiness to adopt, willingness to pay, and trust in a financial AI consultant. We compare the differences over time and in various key situations using a unique experimental framework where participants play a web-based game with real monetary consequences. We observed that accuracy-based explanations of the model in initial phases leads to higher adoption rates. When the performance of the model is immaculate, there is less importance associated with the kind of explanation for adoption. Using more elaborate feature-based or accuracy-based explanations helps substantially in reducing the adoption drop upon model failure. Furthermore, using an autopilot increases adoption significantly. Participants assigned to the AI-labeled advice with explanations were willing to pay more for the advice than the AI-labeled advice with "No-explanation" alternative. These results add to the literature on the importance of XAI for algorithmic adoption and trust.
{"title":"Explainable AI and Adoption of Financial Algorithmic Advisors: An Experimental Study","authors":"D. David, Yehezkel S. Resheff, Talia Tron","doi":"10.1145/3461702.3462565","DOIUrl":"https://doi.org/10.1145/3461702.3462565","url":null,"abstract":"We study whether receiving advice from either a human or algorithmic advisor, accompanied by five types of Local and Global explanation labelings, has an effect on the readiness to adopt, willingness to pay, and trust in a financial AI consultant. We compare the differences over time and in various key situations using a unique experimental framework where participants play a web-based game with real monetary consequences. We observed that accuracy-based explanations of the model in initial phases leads to higher adoption rates. When the performance of the model is immaculate, there is less importance associated with the kind of explanation for adoption. Using more elaborate feature-based or accuracy-based explanations helps substantially in reducing the adoption drop upon model failure. Furthermore, using an autopilot increases adoption significantly. Participants assigned to the AI-labeled advice with explanations were willing to pay more for the advice than the AI-labeled advice with \"No-explanation\" alternative. These results add to the literature on the importance of XAI for algorithmic adoption and trust.","PeriodicalId":197336,"journal":{"name":"Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-01-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125529014","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Govardana Sachithanandam Ramachandran, Ivan Brugere, L. Varshney, Caiming Xiong
Disparate access to resources by different subpopulations is a prevalent issue in societal and sociotechnical networks. For example, urban infrastructure networks may enable certain racial groups to more easily access resources such as high-quality schools, grocery stores, and polling places. Similarly, social networks within universities and organizations may enable certain groups to more easily access people with valuable information or influence. Here we introduce a new class of problems, Graph Augmentation for Equitable Access (GAEA), to enhance equity in networked systems by editing graph edges under budget constraints. We prove such problems are NP-hard, and cannot be approximated within a factor of (1-1/3e). We develop a principled, sample- and time- efficient Markov Reward Process (MRP)-based mechanism design framework for GAEA. Our algorithm outperforms baselines on a diverse set of synthetic graphs. We further demonstrate the method on real-world networks, by merging public census, school, and transportation datasets for the city of Chicago and applying our algorithm to find human-interpretable edits to the bus network that enhance equitable access to high-quality schools across racial groups. Further experiments on Facebook networks of universities yield sets of new social connections that would increase equitable access to certain attributed nodes across gender groups.
{"title":"GAEA: Graph Augmentation for Equitable Access via Reinforcement Learning","authors":"Govardana Sachithanandam Ramachandran, Ivan Brugere, L. Varshney, Caiming Xiong","doi":"10.1145/3461702.3462615","DOIUrl":"https://doi.org/10.1145/3461702.3462615","url":null,"abstract":"Disparate access to resources by different subpopulations is a prevalent issue in societal and sociotechnical networks. For example, urban infrastructure networks may enable certain racial groups to more easily access resources such as high-quality schools, grocery stores, and polling places. Similarly, social networks within universities and organizations may enable certain groups to more easily access people with valuable information or influence. Here we introduce a new class of problems, Graph Augmentation for Equitable Access (GAEA), to enhance equity in networked systems by editing graph edges under budget constraints. We prove such problems are NP-hard, and cannot be approximated within a factor of (1-1/3e). We develop a principled, sample- and time- efficient Markov Reward Process (MRP)-based mechanism design framework for GAEA. Our algorithm outperforms baselines on a diverse set of synthetic graphs. We further demonstrate the method on real-world networks, by merging public census, school, and transportation datasets for the city of Chicago and applying our algorithm to find human-interpretable edits to the bus network that enhance equitable access to high-quality schools across racial groups. Further experiments on Facebook networks of universities yield sets of new social connections that would increase equitable access to certain attributed nodes across gender groups.","PeriodicalId":197336,"journal":{"name":"Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132160052","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Fairness and Outlier Detection (OD) are closely related, as it is exactly the goal of OD to spot rare, minority samples in a given population. However, when being a minority (as defined by protected variables, such as race/ethnicity/sex/age) does not reflect positive-class membership (such as criminal/fraud), OD produces unjust outcomes. Surprisingly, fairness-aware OD has been almost untouched in prior work, as fair machine learning literature mainly focuses on supervised settings. Our work aims to bridge this gap. Specifically, we develop desiderata capturing well-motivated fairness criteria for OD, and systematically formalize the fair OD problem. Further, guided by our desiderata, we propose FairOD, a fairness-aware outlier detector that has the following desirable properties: FairOD (1) exhibits treatment parity at test time, (2) aims to flag equal proportions of samples from all groups (i.e. obtain group fairness, via statistical parity), and (3) strives to flag truly high-risk samples within each group. Extensive experiments on a diverse set of synthetic and real world datasets show that FairOD produces outcomes that are fair with respect to protected variables, while performing comparable to (and in some cases, even better than) fairness-agnostic detectors in terms of detection performance.
{"title":"FairOD: Fairness-aware Outlier Detection","authors":"Shubhranshu Shekhar, Neil Shah, L. Akoglu","doi":"10.1145/3461702.3462517","DOIUrl":"https://doi.org/10.1145/3461702.3462517","url":null,"abstract":"Fairness and Outlier Detection (OD) are closely related, as it is exactly the goal of OD to spot rare, minority samples in a given population. However, when being a minority (as defined by protected variables, such as race/ethnicity/sex/age) does not reflect positive-class membership (such as criminal/fraud), OD produces unjust outcomes. Surprisingly, fairness-aware OD has been almost untouched in prior work, as fair machine learning literature mainly focuses on supervised settings. Our work aims to bridge this gap. Specifically, we develop desiderata capturing well-motivated fairness criteria for OD, and systematically formalize the fair OD problem. Further, guided by our desiderata, we propose FairOD, a fairness-aware outlier detector that has the following desirable properties: FairOD (1) exhibits treatment parity at test time, (2) aims to flag equal proportions of samples from all groups (i.e. obtain group fairness, via statistical parity), and (3) strives to flag truly high-risk samples within each group. Extensive experiments on a diverse set of synthetic and real world datasets show that FairOD produces outcomes that are fair with respect to protected variables, while performing comparable to (and in some cases, even better than) fairness-agnostic detectors in terms of detection performance.","PeriodicalId":197336,"journal":{"name":"Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115275752","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Ranking algorithms are being widely employed in various online hiring platforms including LinkedIn, TaskRabbit, and Fiverr. Prior research has demonstrated that ranking algorithms employed by these platforms are prone to a variety of undesirable biases, leading to the proposal of fair ranking algorithms (e.g., Det-Greedy) which increase exposure of underrepresented candidates. However, there is little to no work that explores whether fair ranking algorithms actually improve real world outcomes (e.g., hiring decisions) for underrepresented groups. Furthermore, there is no clear understanding as to how other factors (e.g., job context, inherent biases of the employers) may impact the efficacy of fair ranking in practice. In this work, we analyze various sources of gender biases in online hiring platforms, including the job context and inherent biases of employers and establish how these factors interact with ranking algorithms to affect hiring decisions. To the best of our knowledge, this work makes the first attempt at studying the interplay between the aforementioned factors in the context of online hiring. We carry out a large-scale user study simulating online hiring scenarios with data from TaskRabbit, a popular online freelancing site. Our results demonstrate that while fair ranking algorithms generally improve the selection rates of underrepresented minorities, their effectiveness relies heavily on the job contexts and candidate profiles.
{"title":"Does Fair Ranking Improve Minority Outcomes? Understanding the Interplay of Human and Algorithmic Biases in Online Hiring","authors":"Tom Sühr, Sophie Hilgard, Himabindu Lakkaraju","doi":"10.1145/3461702.3462602","DOIUrl":"https://doi.org/10.1145/3461702.3462602","url":null,"abstract":"Ranking algorithms are being widely employed in various online hiring platforms including LinkedIn, TaskRabbit, and Fiverr. Prior research has demonstrated that ranking algorithms employed by these platforms are prone to a variety of undesirable biases, leading to the proposal of fair ranking algorithms (e.g., Det-Greedy) which increase exposure of underrepresented candidates. However, there is little to no work that explores whether fair ranking algorithms actually improve real world outcomes (e.g., hiring decisions) for underrepresented groups. Furthermore, there is no clear understanding as to how other factors (e.g., job context, inherent biases of the employers) may impact the efficacy of fair ranking in practice. In this work, we analyze various sources of gender biases in online hiring platforms, including the job context and inherent biases of employers and establish how these factors interact with ranking algorithms to affect hiring decisions. To the best of our knowledge, this work makes the first attempt at studying the interplay between the aforementioned factors in the context of online hiring. We carry out a large-scale user study simulating online hiring scenarios with data from TaskRabbit, a popular online freelancing site. Our results demonstrate that while fair ranking algorithms generally improve the selection rates of underrepresented minorities, their effectiveness relies heavily on the job contexts and candidate profiles.","PeriodicalId":197336,"journal":{"name":"Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133320480","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Umang Bhatt, Yunfeng Zhang, Javier Antorán, Q. Liao, P. Sattigeri, Riccardo Fogliato, Gabrielle Gauthier Melançon, R. Krishnan, Jason Stanley, Omesh Tickoo, L. Nachman, R. Chunara, Adrian Weller, Alice Xiang
Algorithmic transparency entails exposing system properties to various stakeholders for purposes that include understanding, improving, and contesting predictions. Until now, most research into algorithmic transparency has predominantly focused on explainability. Explainability attempts to provide reasons for a machine learning model's behavior to stakeholders. However, understanding a model's specific behavior alone might not be enough for stakeholders to gauge whether the model is wrong or lacks sufficient knowledge to solve the task at hand. In this paper, we argue for considering a complementary form of transparency by estimating and communicating the uncertainty associated with model predictions. First, we discuss methods for assessing uncertainty. Then, we characterize how uncertainty can be used to mitigate model unfairness, augment decision-making, and build trustworthy systems. Finally, we outline methods for displaying uncertainty to stakeholders and recommend how to collect information required for incorporating uncertainty into existing ML pipelines. This work constitutes an interdisciplinary review drawn from literature spanning machine learning, visualization/HCI, design, decision-making, and fairness. We aim to encourage researchers and practitioners to measure, communicate, and use uncertainty as a form of transparency.
{"title":"Uncertainty as a Form of Transparency: Measuring, Communicating, and Using Uncertainty","authors":"Umang Bhatt, Yunfeng Zhang, Javier Antorán, Q. Liao, P. Sattigeri, Riccardo Fogliato, Gabrielle Gauthier Melançon, R. Krishnan, Jason Stanley, Omesh Tickoo, L. Nachman, R. Chunara, Adrian Weller, Alice Xiang","doi":"10.1145/3461702.3462571","DOIUrl":"https://doi.org/10.1145/3461702.3462571","url":null,"abstract":"Algorithmic transparency entails exposing system properties to various stakeholders for purposes that include understanding, improving, and contesting predictions. Until now, most research into algorithmic transparency has predominantly focused on explainability. Explainability attempts to provide reasons for a machine learning model's behavior to stakeholders. However, understanding a model's specific behavior alone might not be enough for stakeholders to gauge whether the model is wrong or lacks sufficient knowledge to solve the task at hand. In this paper, we argue for considering a complementary form of transparency by estimating and communicating the uncertainty associated with model predictions. First, we discuss methods for assessing uncertainty. Then, we characterize how uncertainty can be used to mitigate model unfairness, augment decision-making, and build trustworthy systems. Finally, we outline methods for displaying uncertainty to stakeholders and recommend how to collect information required for incorporating uncertainty into existing ML pipelines. This work constitutes an interdisciplinary review drawn from literature spanning machine learning, visualization/HCI, design, decision-making, and fairness. We aim to encourage researchers and practitioners to measure, communicate, and use uncertainty as a form of transparency.","PeriodicalId":197336,"journal":{"name":"Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126699620","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
R. Mothilal, Divyat Mahajan, Chenhao Tan, Amit Sharma
Feature attributions and counterfactual explanations are popular approaches to explain a ML model. The former assigns an importance score to each input feature, while the latter provides input examples with minimal changes to alter the model's predictions. To unify these approaches, we provide an interpretation based on the actual causality framework and present two key results in terms of their use. First, we present a method to generate feature attribution explanations from a set of counterfactual examples. These feature attributions convey how important a feature is to changing the classification outcome of a model, especially on whether a subset of features is necessary and/or sufficient for that change, which attribution-based methods are unable to provide. Second, we show how counterfactual examples can be used to evaluate the goodness of an attribution-based explanation in terms of its necessity and sufficiency. As a result, we highlight the complimentary of these two approaches. Our evaluation on three benchmark datasets --- Adult-Income, LendingClub, and German-Credit --- confirms the complimentary. Feature attribution methods like LIME and SHAP and counterfactual explanation methods like Wachter et al. and DiCE often do not agree on feature importance rankings. In addition, by restricting the features that can be modified for generating counterfactual examples, we find that the top-k features from LIME or SHAP are often neither necessary nor sufficient explanations of a model's prediction. Finally, we present a case study of different explanation methods on a real-world hospital triage problem.
特征归因和反事实解释是解释ML模型的常用方法。前者为每个输入特征分配一个重要分数，而后者提供最小变化的输入示例来改变模型的预测。为了统一这些方法，我们提供了一个基于实际因果关系框架的解释，并就其使用提出了两个关键结果。首先，我们提出了一种从一组反事实示例中生成特征归因解释的方法。这些特征属性传达了特征对于改变模型的分类结果有多重要，特别是关于特征子集对于该变化是否必要和/或充分，这是基于属性的方法无法提供的。其次，我们展示了如何使用反事实例子来评估基于归因的解释的必要性和充分性。因此，我们强调这两种方法的互补性。我们对三个基准数据集(Adult-Income, LendingClub和German-Credit)的评估证实了这一点。特征归因方法(如LIME和SHAP)和反事实解释方法(如Wachter et al.和DiCE)在特征重要性排名上往往不一致。此外，通过限制可以修改以生成反事实示例的特征，我们发现来自LIME或SHAP的top-k特征通常既不是模型预测的必要解释，也不是充分解释。最后，我们提出了一个案例研究不同的解释方法对现实世界的医院分诊问题。
{"title":"Towards Unifying Feature Attribution and Counterfactual Explanations: Different Means to the Same End","authors":"R. Mothilal, Divyat Mahajan, Chenhao Tan, Amit Sharma","doi":"10.1145/3461702.3462597","DOIUrl":"https://doi.org/10.1145/3461702.3462597","url":null,"abstract":"Feature attributions and counterfactual explanations are popular approaches to explain a ML model. The former assigns an importance score to each input feature, while the latter provides input examples with minimal changes to alter the model's predictions. To unify these approaches, we provide an interpretation based on the actual causality framework and present two key results in terms of their use. First, we present a method to generate feature attribution explanations from a set of counterfactual examples. These feature attributions convey how important a feature is to changing the classification outcome of a model, especially on whether a subset of features is necessary and/or sufficient for that change, which attribution-based methods are unable to provide. Second, we show how counterfactual examples can be used to evaluate the goodness of an attribution-based explanation in terms of its necessity and sufficiency. As a result, we highlight the complimentary of these two approaches. Our evaluation on three benchmark datasets --- Adult-Income, LendingClub, and German-Credit --- confirms the complimentary. Feature attribution methods like LIME and SHAP and counterfactual explanation methods like Wachter et al. and DiCE often do not agree on feature importance rankings. In addition, by restricting the features that can be modified for generating counterfactual examples, we find that the top-k features from LIME or SHAP are often neither necessary nor sufficient explanations of a model's prediction. Finally, we present a case study of different explanation methods on a real-world hospital triage problem.","PeriodicalId":197336,"journal":{"name":"Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116798208","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Typically, fair machine learning research focuses on a single decision maker and assumes that the underlying population is stationary. However, many of the critical domains motivating this work are characterized by competitive marketplaces with many decision makers. Realistically, we might expect only a subset of them to adopt any non-compulsory fairness-conscious policy, a situation that political philosophers call partial compliance. This possibility raises important questions: how does partial compliance and the consequent strategic behavior of decision subjects affect the allocation outcomes? If k% of employers were to voluntarily adopt a fairness-promoting intervention, should we expect k% progress (in aggregate) towards the benefits of universal adoption, or will the dynamics of partial compliance wash out the hoped-for benefits? How might adopting a global (versus local) perspective impact the conclusions of an auditor? In this paper, we propose a simple model of an employment market, leveraging simulation as a tool to explore the impact of both interaction effects and incentive effects on outcomes and auditing metrics. Our key findings are that at equilibrium: (1) partial compliance by k% of employers can result in far less than proportional (k%) progress towards the full compliance outcomes; (2) the gap is more severe when fair employers match global (vs local) statistics; (3) choices of local vs global statistics can paint dramatically different pictures of the performance vis-a-vis fairness desiderata of compliant versus non-compliant employers; (4) partial compliance based on local parity measures can induce extreme segregation. Finally, we discuss implications for auditors and insights concerning the design of regulatory frameworks.
{"title":"Fair Machine Learning Under Partial Compliance","authors":"Jessica Dai, S. Fazelpour, Zachary Chase Lipton","doi":"10.1145/3461702.3462521","DOIUrl":"https://doi.org/10.1145/3461702.3462521","url":null,"abstract":"Typically, fair machine learning research focuses on a single decision maker and assumes that the underlying population is stationary. However, many of the critical domains motivating this work are characterized by competitive marketplaces with many decision makers. Realistically, we might expect only a subset of them to adopt any non-compulsory fairness-conscious policy, a situation that political philosophers call partial compliance. This possibility raises important questions: how does partial compliance and the consequent strategic behavior of decision subjects affect the allocation outcomes? If k% of employers were to voluntarily adopt a fairness-promoting intervention, should we expect k% progress (in aggregate) towards the benefits of universal adoption, or will the dynamics of partial compliance wash out the hoped-for benefits? How might adopting a global (versus local) perspective impact the conclusions of an auditor? In this paper, we propose a simple model of an employment market, leveraging simulation as a tool to explore the impact of both interaction effects and incentive effects on outcomes and auditing metrics. Our key findings are that at equilibrium: (1) partial compliance by k% of employers can result in far less than proportional (k%) progress towards the full compliance outcomes; (2) the gap is more severe when fair employers match global (vs local) statistics; (3) choices of local vs global statistics can paint dramatically different pictures of the performance vis-a-vis fairness desiderata of compliant versus non-compliant employers; (4) partial compliance based on local parity measures can induce extreme segregation. Finally, we discuss implications for auditors and insights concerning the design of regulatory frameworks.","PeriodicalId":197336,"journal":{"name":"Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124633682","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Algorithmic systems are known to impact marginalized groups severely, and more so, if all sources of bias are not considered. While work in algorithmic fairness to-date has primarily focused on addressing discrimination due to individually linked attributes, social science research elucidates how some properties we link to individuals can be conceptualized as having causes at macro (e.g. structural) levels, and it may be important to be fair to attributes at multiple levels. For example, instead of simply considering race as a causal, protected attribute of an individual, the cause may be distilled as perceived racial discrimination an individual experiences, which in turn can be affected by neighborhood-level factors. This multi-level conceptualization is relevant to questions of fairness, as it may not only be important to take into account if the individual belonged to another demographic group, but also if the individual received advantaged treatment at the macro-level. In this paper, we formalize the problem of multi-level fairness using tools from causal inference in a manner that allows one to assess and account for effects of sensitive attributes at multiple levels. We show importance of the problem by illustrating residual unfairness if macro-level sensitive attributes are not accounted for, or included without accounting for their multi-level nature. Further, in the context of a real-world task of predicting income based on macro and individual-level attributes, we demonstrate an approach for mitigating unfairness, a result of multi-level sensitive attributes.
{"title":"Causal Multi-level Fairness","authors":"V. Mhasawade, R. Chunara","doi":"10.1145/3461702.3462587","DOIUrl":"https://doi.org/10.1145/3461702.3462587","url":null,"abstract":"Algorithmic systems are known to impact marginalized groups severely, and more so, if all sources of bias are not considered. While work in algorithmic fairness to-date has primarily focused on addressing discrimination due to individually linked attributes, social science research elucidates how some properties we link to individuals can be conceptualized as having causes at macro (e.g. structural) levels, and it may be important to be fair to attributes at multiple levels. For example, instead of simply considering race as a causal, protected attribute of an individual, the cause may be distilled as perceived racial discrimination an individual experiences, which in turn can be affected by neighborhood-level factors. This multi-level conceptualization is relevant to questions of fairness, as it may not only be important to take into account if the individual belonged to another demographic group, but also if the individual received advantaged treatment at the macro-level. In this paper, we formalize the problem of multi-level fairness using tools from causal inference in a manner that allows one to assess and account for effects of sensitive attributes at multiple levels. We show importance of the problem by illustrating residual unfairness if macro-level sensitive attributes are not accounted for, or included without accounting for their multi-level nature. Further, in the context of a real-world task of predicting income based on macro and individual-level attributes, we demonstrate an approach for mitigating unfairness, a result of multi-level sensitive attributes.","PeriodicalId":197336,"journal":{"name":"Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134255873","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Shubham Sharma, Alan H. Gee, D. Paydarfar, J. Ghosh
Fairness and robustness in machine learning are crucial when individuals are subject to automated decisions made by models in high-stake domains. To promote ethical artificial intelligence, fairness metrics that rely on comparing model error rates across subpopulations have been widely investigated for the detection and mitigation of bias. However, fairness measures that rely on comparing the ability to achieve recourse have been relatively unexplored. In this paper, we present a novel formulation for training neural networks that considers the distance of data observations to the decision boundary such that the new objective: (1) reduces the disparity in the average ability of recourse between individuals in each protected group, and (2) increases the average distance of data points to the boundary to promote adversarial robustness. We demonstrate that models trained with this new objective are more fair and adversarially robust neural networks, with similar accuracies, when compared to models without it. We also investigate a trade-off between the recourse-based fairness and robustness objectives. Moreover, we qualitatively motivate and empirically show that reducing recourse disparity across protected groups also improves fairness measures that rely on error rates. To the best of our knowledge, this is the first time that recourse disparity across groups are considered to train fairer neural networks.
{"title":"FaiR-N: Fair and Robust Neural Networks for Structured Data","authors":"Shubham Sharma, Alan H. Gee, D. Paydarfar, J. Ghosh","doi":"10.1145/3461702.3462559","DOIUrl":"https://doi.org/10.1145/3461702.3462559","url":null,"abstract":"Fairness and robustness in machine learning are crucial when individuals are subject to automated decisions made by models in high-stake domains. To promote ethical artificial intelligence, fairness metrics that rely on comparing model error rates across subpopulations have been widely investigated for the detection and mitigation of bias. However, fairness measures that rely on comparing the ability to achieve recourse have been relatively unexplored. In this paper, we present a novel formulation for training neural networks that considers the distance of data observations to the decision boundary such that the new objective: (1) reduces the disparity in the average ability of recourse between individuals in each protected group, and (2) increases the average distance of data points to the boundary to promote adversarial robustness. We demonstrate that models trained with this new objective are more fair and adversarially robust neural networks, with similar accuracies, when compared to models without it. We also investigate a trade-off between the recourse-based fairness and robustness objectives. Moreover, we qualitatively motivate and empirically show that reducing recourse disparity across protected groups also improves fairness measures that rely on error rates. To the best of our knowledge, this is the first time that recourse disparity across groups are considered to train fairer neural networks.","PeriodicalId":197336,"journal":{"name":"Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128029325","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：