DeepAI AI Chat
Log In Sign Up

It's easy to fool yourself: Case studies on identifying bias and confounding in bio-medical datasets

by   Subhashini Venugopalan, et al.

Confounding variables are a well known source of nuisance in biomedical studies. They present an even greater challenge when we combine them with black-box machine learning techniques that operate on raw data. This work presents two case studies. In one, we discovered biases arising from systematic errors in the data generation process. In the other, we found a spurious source of signal unrelated to the prediction task at hand. In both cases, our prediction models performed well but under careful examination hidden confounders and biases were revealed. These are cautionary tales on the limits of using machine learning techniques on raw data from scientific experiments.


page 2

page 3

page 4


Bias-Resilient Neural Network

Presence of bias and confounding effects is inarguably one of the most c...

Measuring the effects of confounders in medical supervised classification problems: the Confounding Index (CI)

Over the years, there has been growing interest in using Machine Learnin...

Machine Learning for Clinical Predictive Analytics

In this chapter, we provide a brief overview of applying machine learnin...

Pulling Up by the Causal Bootstraps: Causal Data Augmentation for Pre-training Debiasing

Machine learning models achieve state-of-the-art performance on many sup...

Quantifying Confounding Bias in Neuroimaging Datasets with Causal Inference

Neuroimaging datasets keep growing in size to address increasingly compl...

Identifying the Context Shift between Test Benchmarks and Production Data

Across a wide variety of domains, there exists a performance gap between...

Hybrid Method Based on NARX models and Machine Learning for Pattern Recognition

This work presents a novel technique that integrates the methodologies o...