July 03, 2019

Mammography Challenge Enables Residents Without Coding Skills to Create AI

Coming this August, ACR AI-LAB™ will host an AI Resident Mammography Challenge. Residents in the challenge will train a model using annotated data they create themselves and tune the model by adjusting hyperparameters with ACR AI-LAB. Unlike in most AI challenges, coding skills will not be needed for AI newcomers to participate.

The challenge is designed to educate residents about the essential skills involved in AI, including annotating data, building models and testing models for performance.

At the start of the challenge, residents will receive a model pre-trained with ImageNet to classify natural images, such as cats, dogs, horses and trees. From there, residents will further train the model to classify breast density studies by taking a set of data they receive to first annotate data and then to create training and validation data sets.

The resident who submits the model with the best kappa score will win compensation for attendance at RSNA 2019. Look for more information on entering the AI Resident Mammography Challenge on ACR AI-LAB.