top of page

MODEL MADNESS

MODEL MADNESS is an offline competition where students develop machine learning models to solve medical industry challenges. Given an image dataset, participants must build and optimize their models within a set time, applying deep learning and computer vision. The team with the best performance wins.

Computer Circuit Board

RULES & REGULATIONS

Event Format:

  • Platform: Kaggle (using the P100 GPU)

  • Dataset: A healthcare-related image dataset will be provided at the start of the competition.

  • Task: Teams must build and train a machine learning model using the dataset and optimize it based on a performance metric that will be revealed on the day of the event.

  • Evaluation Metric: This will be disclosed during the competition. Teams will be evaluated based on their model's performance on this metric.

Participation Details:

  • Team Size: 2 participants per team

  • Eligibility: Open to students from Grades 9 to 12

  • Entry Limit: One team per school

  • Event Duration: 3 hours

Judgment Criteria:

  • The winning team will be determined by the maximum performance metric achieved by their machine learning model on the evaluation dataset.

Requirements:

  • Participants should have experience in deep learning and computer vision, particularly with Convolutional Neural Networks (CNNs).

  • Familiarity with standard machine learning libraries, such as TensorFlow, PyTorch, and Scikit-learn, is essential.

  • Participants must have a Kaggle account with at least 5 hours of GPU quota available for the competition day.

  • Teams are encouraged to prepare in advance by reviewing model evaluation techniques and performance optimization strategies.

image.png

DPSI TECHATHLON 2024

HS-01, Golf Course Ext Rd, Block W, South City II, Sector 50, Gurugram, Haryana 122001

bottom of page