Scientific Abstract | Fertility and Sterility

Identifying potential sources of bias in deep learning models for embryo assessment

October 19th, 2021

Abstract

Objective: To identify and reduce potential sources of bias when training deep learning models for analyzing images of human embryos.

Impact Statement: Naive approaches to preparing training data for deep learning models for embryo ranking can create bias in the models. Our work illustrates the need for careful preparation of training data and monitoring of different metrics to identify and reduce potential sources of bias.