skip to main content

Attention:

The NSF Public Access Repository (PAR) system and access will be unavailable from 11:00 PM ET on Thursday, February 13 until 2:00 AM ET on Friday, February 14 due to maintenance. We apologize for the inconvenience.


Search for: All records

Creators/Authors contains: "Schelter, Sebastian"

Note: When clicking on a Digital Object Identifier (DOI) number, you will be taken to an external site maintained by the publisher. Some full text articles may not yet be available without a charge during the embargo (administrative interval).
What is a DOI Number?

Some links on this page may take you to non-federal websites. Their policies may differ from this site.

  1. In this paper, we interrogate whether data quality issues track demographic group membership (based on sex, race and age) and whether automated data cleaning — of the kind commonly used in production ML systems — impacts the fairness of predictions made by these systems. To the best of our knowledge, the impact of data cleaning on fairness in downstream tasks has not been investigated in the literature. We first analyse the tuples flagged by common error detection strategies in five research datasets. We find that, while specific data quality issues, such as higher rates of missing values, are associated with membership in historically disadvantaged groups, poor data quality does not generally track demographic group membership. As a follow-up, we conduct a large-scale empirical study on the impact of automated data cleaning on fairness, involving more than 26,000 model evaluations. We observe that, while automated data cleaning is unlikely to worsen accuracy, it is more likely to worsen fairness than to improve it, especially when the cleaning techniques are not carefully chosen. Furthermore, we find that the positive or negative impact of a particular cleaning technique often depends on the choice of fairness metric and group definition (single-attribute or intersectional). We make our code and experimental results publicly available. The analysis we conducted in this paper is difficult, primarily because it requires that we think holistically about disparities in data quality, disparities in the effectiveness of data cleaning methods, and impacts of such disparities on ML model performance for different demographic groups. Such holistic analysis can and should be supported by data engineering tools, and requires substantial data engineering research. Towards this goal, we discuss open research questions, envision the development of fairness-aware data cleaning methods, and their integration into complex pipelines for ML-based decision making. 
    more » « less
  2. In this paper, we interrogate whether data quality issues track demographic characteristics such as sex, race and age, and whether automated data cleaning — of the kind commonly used in production ML systems — impacts the fairness of predictions made by these systems. To the best of our knowledge, the impact of data cleaning on fairness in downstream tasks has not been investigated in the literature.We first analyze the tuples flagged by common error detection strategies in five research datasets. We find that, while specific data quality issues, such as higher rates of missing values, are associated with membership in historically disadvantaged groups, poor data quality does not generally track demographic group membership. As a follow-up, we conduct a large-scale empirical study on the impact of automated data cleaning on fairness, involving more than 26,000 model evaluations on five datasets. We observe that, while automated data cleaning has an insignificant impact on both accuracy and fairness in the majority of cases, it is more likely to worsen fairness than to improve it, especially when the cleaning techniques are not carefully chosen. This finding is both significant and worrying, given that it potentially implicates many production ML systems. We make our code and experimental results publicly available.The analysis we conducted in this paper is difficult, primarily because it requires that we think holistically about disparities in data quality, disparities in the effectiveness of data cleaning methods, and impacts of such disparities on ML model performance for different demographic groups. Such holistic analysis can and should be supported with the help of data engineering research. Towards this goal, we envision the development of fairness-aware data cleaning methods, and their integration into complex pipelines for ML-based decision making. 
    more » « less
  3. Perspectives on the role and responsibility of the data-management research community in designing, developing, using, and overseeing automated decision systems. 
    more » « less
  4. Machine learning (ML) is increasingly used to automate impactful decisions, and the risks arising from this widespread use are garnering attention from policy makers, scientists, and the media. ML applications are often brittle with respect to their input data, which leads to concerns about their correctness, reliability, and fairness. In this paper, we describe mlinspect, a library that helps diagnose and mitigate technical bias that may arise during preprocessing steps in an ML pipeline. We refer to these problems collectively as data distribution bugs. The key idea is to extract a directed acyclic graph representation of the dataflow from a preprocessing pipeline and to use this representation to automatically instrument the code with predefined inspections. These inspections are based on a lightweight annotation propagation approach to propagate metadata such as lineage information from operator to operator. In contrast to existing work, mlinspect operates on declarative abstractions of popular data science libraries like estimator/transformer pipelines and does not require manual code instrumentation. We discuss the design and implementation of the mlinspect library and give a comprehensive end-to-end example that illustrates its functionality. 
    more » « less
  5. Foulds, James ; Pan, Shimei (Ed.)
    Machine Learning (ML) is commonly used to automate decisions in domains as varied as credit and lending, medical diagnosis, and hiring. These decisions are consequential, imploring us to carefully balance the benefits of efficiency with the potential risks. Much of the conversation about the risks centers around bias — a term that is used by the technical community ever more frequently but that is still poorly understood. In this paper we focus on technical bias — a type of bias that has so far received limited attention and that the data engineering community is well-equipped to address. We discuss dimensions of technical bias that can arise through the ML lifecycle, particularly when it’s due to preprocessing decisions or post-deployment issues. We present results of our recent work, and discuss future research directions. Our over-all goal is to support the development of systems that expose the knobs of responsibility to data scientists, allowing them to detect instances of technical bias and to mitigate it when possible. 
    more » « less
  6. null (Ed.)
    Machine Learning (ML) is increasingly used to automate impactful decisions, and the risks arising from this wide-spread use are garnering attention from policy makers, scientists, and the media. ML applications are often very brittle with respect to their input data, which leads to concerns about their reliability, accountability, and fairness. In this paper we discuss such hard-to-identify data issues and describe mlinspect, a library that enables lightweight lineage-based inspection of ML preprocessing pipelines. The key idea is to extract a directed acyclic graph representation of the data flow from ML preprocessing pipelines in Python, and to use this representation to automatically instrument the code with predefined inspections based on a lightweight annotation propagation approach. In contrast to existing work, mlinspect operates on declarative abstractions of popular data science libraries like estimator/transformer pipelines and does not require manual code instrumentation. We discuss the design and implementation of the mlinspect prototype, and give a complex end-to-end example that illustrates its functionality. 
    more » « less
  7. null (Ed.)
    Machine Learning (ML) is increasingly used to automate impactful decisions, and the risks arising from this wide-spread use are garnering attention from policymakers, scientists, and the media. ML applications are often very brittle with respect to their input data, which leads to concerns about their reliability, accountability, and fairness. While bias detection cannot be fully automated, computational tools can help pinpoint particular types of data issues. We recently proposed mlinspect, a library that enables lightweight lineage-based inspection of ML preprocessing pipelines. In this demonstration, we show how mlinspect can be used to detect data distribution bugs in a representative pipeline. In contrast to existing work, mlinspect operates on declarative abstractions of popular data science libraries like estimator/transformer pipelines, can handle both relational and matrix data, and does not require manual code instrumentation. The library is publicly available at https://github.com/stefan-grafberger/mlinspect. 
    more » « less
  8. The importance of incorporating ethics and legal compliance into machine-assisted decision-making is broadly recognized. Further, several lines of recent work have argued that critical opportunities for improving data quality and representativeness, controlling for bias, and allowing humans to oversee and impact computational processes are missed if we do not consider the lifecycle stages upstream from model training and deployment. Yet, very little has been done to date to provide system-level support to data scientists who wish to develop responsible machine learning methods. We aim to fill this gap and present FairPrep, a design and evaluation framework for fairness-enhancing interventions, which helps data scientists follow best practices in ML experimentation. We identify shortcomings in existing empirical studies for analyzing fairness-enhancing interventions and show how FairPrep can be used to measure their impact. Our results suggest that the high variability of the outcomes of fairness-enhancing interventions observed in previous studies is often an artifact of a lack of hyperparameter tuning, and that the choice of a data cleaning method can impact the effectiveness of fairness-enhancing interventions 
    more » « less
  9. Surfacing and mitigating bias in ML pipelines is a complex topic, with a dire need to provide system-level support to data scientists. Humans should be empowered to debug these pipelines, in order to control for bias and to improve data quality and representativeness. We propose fairDAGs, an open-source library that extracts directed acyclic graph (DAG) representations of the data flow in preprocessing pipelines for ML. The library subsequently instruments the pipelines with tracing and visualization code to capture changes in data distributions and identify distortions with respect to protected group membership as the data travels through the pipeline. We illustrate the utility of fairDAGs, with experiments on publicly available ML pipelines. 
    more » « less