Challenge Stages

The AI Health Outcomes Challenge will run in three stages:  Launch Stage, Stage 1, and Stage 2.
The Launch Stage is open to all eligible participants and the competition is intended to attract a wide range of ideas and participants (e.g., large and small entities, individuals and teams, etc.)

Launch
STAGE

Winners:

25

Stage Description

The Launch Stage is an entry stage where participation will be open to the general public. Entrants will complete an online application and submit a brief slide deck providing information about the participants and their proposed solution, describing their AI solution, based on the application instructions.

Submission Requirements

  1. Online Application completed on the Challenge Platform
  2. Slide Deck (10 slides max) uploaded to Challenge Platform

Dataset Delivery

None.
Participants use data dictionaries and summary description of datasets to craft proposal.

Stage 1:
DESIGN

Winners:

7 awarded $60,000

Stage Description

Stage 1 is a design stage that asks participants to submit proof-of-concept AI solutions related to the clinical problem description, and an associated white paper describing the proof-of-concept solution, and other required elements of the entry.

Participants must run an analysis using the LDS to demonstrate that their solution works.

Submission Requirements

  1. 1st  White Paper uploaded to Challenge Platform
  2. Proof of Concept Algorithm submitted via CMS’ Github

Dataset Delivery

LDS #1 requested and delivered after DUA signed.
Participants use dataset to develop white paper and Proof of Concept Algorithm.

Stage 2:
Refine

Winners:

1 Grand Prize Winner
awarded 1,000,000

2nd place Winner
awarded $230,000

Stage Description

Stage 2 of the challenge asks finalists to further refine their proof-of-concept algorithm related to the clinical problem description and provide an associated white paper describing the concept and other required elements of the entry.

Additionally, finalists will be required to run a series of analyses using the LDS data to prove that their concept works, demonstrating predictive power.

Submission Requirements

  1. 2nd White Paper uploaded to Challenge Platform
  2. Refined Proof of Concept Algorithm submitted via CMS’ Github

Dataset Delivery

LDS #2 requested and delivered after DUA signed.
Participants use larger dataset to refine the AI solution.

Close Menu