Epistemic values in feature importance methods

Hancox-Li, Leif; Kumar, I. Elizabeth

doi:10.1145/3442188.3445943

Cited by 21 publications

(13 citation statements)

References 52 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The practice of "studying up" in machine learning reverses essential assumptions by turning the lens of machine predictions on the social norms and cultural context of those holding power [1]. In a similar vein to our work, Hancox-Li et al critiqued the limitations of feature importance methods and suggests methodologies from feminist epistemology to address them [20]. In their writing, context sensitivity and interactive ways of knowing agree with our theory of identity as ongoing relational processes.…”

Section: Essential Vs Co-constructedmentioning

confidence: 56%

Subverting machines, fluctuating identities: Re-learning human categorization

Lu,

Kay,

McKee

2022

Preprint

View full text Add to dashboard Cite

Most machine learning systems that interact with humans construct some notion of a person's "identity, " yet the default paradigm in AI research envisions identity with essential attributes that are discrete and static. In stark contrast, strands of thought within critical theory present a conception of identity as malleable and constructed entirely through interaction; a doing rather than a being. In this work, we distill some of these ideas for machine learning practitioners and introduce a theory of identity as autopoiesis, circular processes of formation and function. We argue that the default paradigm of identity used by the field immobilizes existing identity categories and the power differentials that co-occur, due to the absence of iterative feedback to our models. This includes a critique of emergent AI fairness practices that continue to impose the default paradigm. Finally, we apply our theory to sketch approaches to autopoietic identity through multilevel optimization and relational learning. While these ideas raise many open questions, we imagine the possibilities of machines that are capable of expressing human identity as a relationship perpetually in flux.

show abstract

Section: Essential Vs Co-constructedmentioning

confidence: 56%

Subverting machines, fluctuating identities: Re-learning human categorization

Lu,

Kay,

McKee

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Designing EMR interfaces which consider a pluralism of explanations (e.g., assigning greater uncertainty to correlated features or incorporating Bayesian methods) in a less familiar form (e.g., different icons to accentuate less relevant variables and disrupt automatic thinking) may invite critical interpretation of the model findings while remaining sensitive to individuals who may be disempowered in the algorithmic design and interpretation phase ( 67 – 70 ). We point the reader to a few cited studies for those seeking a deeper understanding ( 63 – 65 , 71 – 73 ).…”

Section: Discussionmentioning

confidence: 99%

Machine Learning Approaches for Hospital Acquired Pressure Injuries: A Retrospective Study of Electronic Medical Records

Levy

Lima

Miller

et al. 2022

Front. Med. Technol.

View full text Add to dashboard Cite

BackgroundMany machine learning heuristics integrate well with Electronic Medical Record (EMR) systems yet often fail to surpass traditional statistical models for biomedical applications.ObjectiveWe sought to compare predictive performances of 12 machine learning and traditional statistical techniques to predict the occurrence of Hospital Acquired Pressure Injuries (HAPI).MethodsEMR information was collected from 57,227 hospitalizations acquired from Dartmouth Hitchcock Medical Center (April 2011 to December 2016). Twelve classification algorithms, chosen based upon classic regression and recent machine learning techniques, were trained to predict HAPI incidence and performance was assessed using the Area Under the Receiver Operating Characteristic Curve (AUC).ResultsLogistic regression achieved a performance (AUC = 0.91 ± 0.034) comparable to the other machine learning approaches. We report discordance between machine learning derived predictors compared to the traditional statistical model. We visually assessed important patient-specific factors through Shapley Additive Explanations.ConclusionsMachine learning models will continue to inform clinical decision-making processes but should be compared to traditional modeling approaches to ensure proper utilization. Disagreements between important predictors found by traditional and machine learning modeling approaches can potentially confuse clinicians and need to be reconciled. These developments represent important steps forward in developing real-time predictive models that can be integrated into EMR systems to reduce unnecessary harm.

show abstract

“…Other work has considered the implications of feminist epistemologies to areas of ML. For example, Hancox-Li and Kumar [42] apply the frameworks of situated knowledge and standpoint theory [44] to understand the values implicit in feature importance methods. Barabas et al [4] explore how the concept of "studying up" [59] could be used to reorient ML research questions to better confront power.…”

Section: Participatory and Feminist MLmentioning

confidence: 99%

Towards Intersectional Feminist and Participatory ML: A Case Study in Supporting Feminicide Counterdata Collection

Suresh

Rajiv

Dogan

et al. 2022

2022 ACM Conference on Fairness, Accountability, and Transparency

View full text Add to dashboard Cite

Data ethics and fairness have emerged as important areas of research in recent years. However, much work in this area focuses on retroactively auditing and "mitigating bias" in existing, potentially flawed systems, without interrogating the deeper structural inequalities underlying them. There are not yet examples of how to apply feminist and participatory methodologies from the start, to conceptualize and design machine learning-based tools that center and aim to challenge power inequalities. Our work targets this more prospective goal. Guided by the framework of data feminism, we co-design datasets and machine learning models to support the efforts of activists who collect and monitor data about feminicide -gender-based killings of women and girls. We describe how intersectional feminist goals and participatory processes shaped each stage of our approach, from problem conceptualization to data collection to model evaluation. We highlight several methodological contributions, including 1) an iterative data collection and annotation process that targets model weaknesses and interrogates framing concepts (such as who is included/excluded in "feminicide"), 2) models that explicitly focus on intersectional identities rather than statistical majorities, and 3) a multi-step evaluation process -with quantitative, qualitative and participatory stepsfocused on context-specific relevance. We also distill insights and tensions that arise from bridging intersectional feminist goals with ML. These include reflections on how ML may challenge power, embrace pluralism, rethink binaries and consider context, as well as the inherent limitations of any technology-based solution to address durable structural inequalities.

show abstract

Epistemic values in feature importance methods

Cited by 21 publications

References 52 publications

Subverting machines, fluctuating identities: Re-learning human categorization

Subverting machines, fluctuating identities: Re-learning human categorization

Machine Learning Approaches for Hospital Acquired Pressure Injuries: A Retrospective Study of Electronic Medical Records

Towards Intersectional Feminist and Participatory ML: A Case Study in Supporting Feminicide Counterdata Collection

Contact Info

Product

Resources

About