The Sorting Hat Project

Written by The Sorting Hat Project Team

Have you ever wondered… “What specialty should I apply into?”

“Am I picking the right specialty for me?”

“I wish someone would just tell me which specialty to apply into!”

If only… there was a way to know…

The goal of the Sorting Hat Project is to train an AI algorithm to predict a medical student’s future specialty choice using only their headshot, date of birth, and photo of their right palm. In order to do this, we need to first collect this dataset of headshots and palmshots, which we have aptly named HoGWARTS - “Hand or Grinning-While-Actually-Really-Tired Snapshot” Dataset.

“But wait- this idea is crazy. Why would you do this? Is there any precedent?”

Well, our main desire is to just see if this crazy idea could work, and to practice our coding with a fun project.

People in the past have done some pretty unexpected things with AI. They’ve been able to predict MI’s using retinal scans [1] and personality traits from video interviews [2]. We are wondering… what can AI see that we cannot? Perhaps our future specialty… from the contours of our hands and faces? 

We would like to ask for current MS4s and PGY’s/MD’s/DO’s to please submit data so we can start training this algorithm! We will only need your ERAS headshot and a photo of your right palm. We are trying to build as diverse of a dataset as possible and would love to have every specialty included. 

If this works, we will make our algorithm freely available to help assuage the existential fears of all future generations of medical students in a light-hearted and playful way. 

Find out more here, at sortinghatproject.com!

Twitter: @SortingHatProj

[1] https://www.nature.com/articles/s42256-021-00427-7

[2] https://ieeexplore.ieee.org/abstract/document/8660507