Can You Fake It Until You Make It?

Cheng, Victoria; Suriyakumar, Vinith M.; Dullerud, Natalie; Joshi, Shalmali; Ghassemi, Marzyeh

doi:10.1145/3442188.3445879

Cited by 28 publications

(14 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This could have practical benefits when data sharing to protect personally identifiable information (PII) while achieving high quality performance. Such an approach would of course be associated with its own risks, some of which are discussed by Cheng et al (2021).…”

Section: Discussionmentioning

confidence: 99%

Generative models improve fairness of medical classifiers under distribution shifts

Wiles

Albuquerque

Rebuffi

et al. 2023

Preprint

View full text Add to dashboard Cite

Domain generalization remains a ubiquitous challenge for machine learning in healthcare. Model performance in real-world conditions might be lower than expected due to discrepancies between the data encountered in deployment environments and datasets used for model development. Under-representation of some groups or conditions during model development is a common cause of this phenomenon, which can have serious implications as it can exacerbate bias against groups, individuals or conditions and propagate unintended harms in their care. This challenge is often not readily addressed by targeted data acquisition and “labelling” by expert clinicians, which can be prohibitively expensive or practically impossible due to the rarity of diseases, conditions, or available clinical expertise. We hypothesize that advances in generative artificial intelligence may help mitigate this unmet need in a steerable fashion, algorithmically enriching our training dataset with synthetic examples that address shortfalls of underrepresented conditions or subgroups. We show that generative models can automatically learn realistic augmentations from data in a label-efficient manner. In particular, we leverage the higher abundance of unlabelled data to model the underlying distribution of different conditions and subgroups for an imaging modality. By conditioning generative models on appropriate labels (e.g., diagnostic labels and / or sensitive attribute labels), we can steer the distribution of synthetic examples according to specific requirements. We demonstrate that these learned augmentations make models more robust and statistically fair in- and out-of-distribution. To evaluate the generality of our approach, we study three distinct medical imaging contexts of varying difficulty: (i) histopathology images from a publicly available and widely adopted generalization benchmark, (ii) chest X-rays from publicly available clinical datasets, and (iii) dermatology images characterized by complex shifts and imaging conditions. The latter constitutes a particularly unstructured domain with various challenges. Two of these imaging modalities further require operating at a high-resolution, which requires developing faithful super-resolution techniques to recover fine details of each health condition. Complementing real training samples with synthetic ones improves the robustness of models in all three medical tasks and increases fairness by improving the accuracy of clinical diagnosis within underrepresented groups. Our proposed approach leads to stark improvements out-of-distribution across modalities: 7.7% prediction accuracy improvement in histopathology, 5.2% in chest radiology with 44.6% lower fairness gap and a striking 63.5% improvement in high-risk sensitivity for dermatology with a 7.5x reduction in fairness gap.

show abstract

Section: Discussionmentioning

confidence: 99%

Generative models improve fairness of medical classifiers under distribution shifts

Wiles

Albuquerque

Rebuffi

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…In the meantime, differential privacy has been shown to amplify the fairness issues in the original data [118]. [119] demonstrate that differential privacy does not introduce unfairness into the data generation process or to standard group fairness measures in the downstream classification models, but does unfairly increase the influence of majority subgroups. Differential privacy also significantly reduces the quality of the images generated from the GANs, which decrease the utility of the synthetic data in downstream tasks.…”

Section: Fairnessmentioning

confidence: 90%

Machine Learning for Synthetic Data Generation: a Review

Lu¹,

Wang²,

Wei³

2023

Preprint

View full text Add to dashboard Cite

Data plays a crucial role in machine learning. However, in real-world applications, there are several problems with data, e.g., data are of low quality; a limited number of data points lead to under-fitting of the machine learning model; it is hard to access the data due to privacy, safety and regulatory concerns. Synthetic data generation offers a promising new avenue, as it can be shared and used in ways that real-world data cannot. This paper systematically reviews the existing works that leverage machine learning models for synthetic data generation. Specifically, we discuss the synthetic data generation works from several perspectives: (i) applications, including computer vision, speech, natural language, healthcare, and business; (ii) machine learning methods, particularly neural network architectures and deep generative models; (iii) privacy and fairness issue. In addition, we identify the challenges and opportunities in this emerging field and suggest future research directions.

show abstract

“…Very recently, synthetic data generation methods have begun to adopt fairness as one of the key elements of equitable data generation. Differentially private GANs attempt to preserve privacy and mitigate fairness but have shown to lead to decreased synthetic image quality and utility [27]. Others measured synthetic data fairness by developing models on the data and then calculated fairness metrics like demographic parity, equality of odds etc.…”

Section: Related Workmentioning

confidence: 99%

“…They too found that models for differentially private data result in more disparity and poorer fairness results. In [27], the authors looked at group fairness metrics like parity gap, specificity gap, recall gap, etc. on a downstream model generated on the synthetic data.…”

Section: Related Workmentioning

confidence: 99%

“…For example, for ATUS, for both gender and age individually, the metric shows the results to be fair. However, when we look at the combination of protected attributes, it becomes apparent that some subgroups are under-represented (female aged 75+) while others are over-represented (males aged [25][26][27][28][29][30][31][32][33][34].…”

Section: Need For Covariate Fairness Metricsmentioning

confidence: 99%

See 1 more Smart Citation

The Problem of Fairness in Synthetic Healthcare Data

Bhanot

Miao

Erickson

et al. 2021

Entropy

View full text Add to dashboard Cite

Access to healthcare data such as electronic health records (EHR) is often restricted by laws established to protect patient privacy. These restrictions hinder the reproducibility of existing results based on private healthcare data and also limit new research. Synthetically-generated healthcare data solve this problem by preserving privacy and enabling researchers and policymakers to drive decisions and methods based on realistic data. Healthcare data can include information about multiple in- and out- patient visits of patients, making it a time-series dataset which is often influenced by protected attributes like age, gender, race etc. The COVID-19 pandemic has exacerbated health inequities, with certain subgroups experiencing poorer outcomes and less access to healthcare. To combat these inequities, synthetic data must “fairly” represent diverse minority subgroups such that the conclusions drawn on synthetic data are correct and the results can be generalized to real data. In this article, we develop two fairness metrics for synthetic data, and analyze all subgroups defined by protected attributes to analyze the bias in three published synthetic research datasets. These covariate-level disparity metrics revealed that synthetic data may not be representative at the univariate and multivariate subgroup-levels and thus, fairness should be addressed when developing data generation methods. We discuss the need for measuring fairness in synthetic healthcare data to enable the development of robust machine learning models to create more equitable synthetic healthcare datasets.

show abstract

Can You Fake It Until You Make It?

Cited by 28 publications

References 26 publications

Generative models improve fairness of medical classifiers under distribution shifts

Generative models improve fairness of medical classifiers under distribution shifts

Machine Learning for Synthetic Data Generation: a Review

The Problem of Fairness in Synthetic Healthcare Data

Contact Info

Product

Resources

About