Call for Abstract (Problem Statement) is now Closed with 50 teams!
Registration is FREE! However, as we can only host a limited number of teams this year, the selected teams will need to pay a small sum of deposit, which will be fully refunded after the team has finished the datathon.
Champion : SGD 5,000 cash + USD 1,500 Cloud Credits
1st Runner-up: SGD 3,000 cash + USD 1,000 Cloud Credits
2nd Runner-up: SGD 2,000 cash + USD 1,000 Cloud Credits
Best presentation: SGD 500 + USD 500 Cloud Credits
Best potential: SGD 500 + USD 500 Cloud Credits
Crowd's favorite: SGD 500 + USD 500 Cloud Credits
Finalists (Top 10): Certification from organizers
Computational resources up to 4XV100 GPUs will be provided by Huawei Cloud !!!
Same as previous years, our healthcare AI datathon is to cultivate collaborations between the healthcare and data science experts. We encourage each cross-disciplinary team to propose a problem statement addressing an actual clinical pain point. Teams are tasks to produce some preliminary results to validate the feasibilities and clinical impact of their proposed solution over the period of the datathon. The organizers will be providing a number of real-world datasets, covering a wide spectrum of healthcare data ranging from Electronic Medical Records, Clinical Text to Medical Images (Details can refer to the Data page). The organizers will also support the teams with certain computational powers through cloud services. The best teams stand a chance to win up to SGD 10,000 cash and credit prizes.
Start: 12th Oct to 30th Nov 2020
- Abstract (Problem Statment) submission with initial team members (if any)
- Data scientists/machine learning experts registration (matching will start once abstracts are selected)
Round 1: 30th Nov to 3rd Dec 2020
Abstract (Problem Statement) Selection: to be announced by 30th Nov 2020
Round 2: 30th Nov 2020 to 6th Dec 2020
- Team forming for teams still recruiting members
- Team member confirmation by 6th Dec 2020
Datathon: 11th to 13th Dec 2020
Final Presentation & Judging: 8pm 13th Dec 2020
The following are the rules for the datathon and team forming:
1. Each team MUST have at least one clinical team leader. The clinical team leader will have to submit the problem statement (abstract) by 16th Nov 2020.
2. All the abstracts will be reviewed by our scientific mentors. Judging based on the novelty and impact of the proposed problem statement, the selected abstracts and teams will be informed by 30th Nov 2020.
3. NOTE: Teams should NOT reuse project problem statements that they have previously published or have been already working on for a long time. Teams are encouraged to propose new problem statements for the datathon. A declaration is needed during registration.
4. Each team will be at least 2 and at most 10 members. For an effective team, we recommend the team should contain a number of healthcare experts, data engineers familiar with SQL to extract data and data scientists that can build predictive or statistical models.
5. Clinical/data science team leaders can submit their abstract with their own team members (fully formed or partially form), or they are welcome to participate as an individual. For individual clinical team leaders, if their abstract is selected, we will match them with registered data scientists to form their team. Teams can no longer be changed after 6th Dec 2020.
6. Data scientists/machine learning experts can register now as well. We will match you to the selected teams that are still looking for members.
7. All team members need to apply and obtain access to the datasets that they intend to use before the start of the datathon. To be fair, teams are only allowed to use the public datasets that the event supports.
8. All teams will need to submit their pre-recorded presentation video (no more than 5 minutes) and upload it on Microsoft Team "Final Video Submission" channel by 7PM 13th Dec (Singapore time). No or late submission will be automatically disqualified!
Clinical Relevance and Impact: how your solution may influence or improve the current healthcare practice
Method Novelty: NOT complexity
Team's Plan for Moving Forward