Ryver.ai, Beat t. med stud, from sklearn import linear_model

Start page of the game

Inspiration

We were shocked by the fact that a technology so certain to have a great impact on the field of radiology is so poorly included in the associated education.

What it does

The game gets radiology trainees in touch with AI technology and gives them an idea of the way an AI model makes a decision.

How we built it

We used a pre-trained dense net with an implementation of a heatmapper. Furthermore we have used the vue.js framework to develop a front end for our quiz and a way to interface with the AI decision.

Challenges we ran into

Finding a proper model architecture to include explainability of the AI decision for non-technical radiologists.

Accomplishments that we're proud of

We have functioned well as team and it took us little effort to feel comfortable and have fun collaborating with each other.

What we learned

Different for everyone of the team. Ranges from first experience in front-end design to implementing ways to scale explainability of AI models in a medical context.