AI for Social Good Project
In the second half of the program, students will work in groups of three to implement a mentor-led socially impactful AI project.
Students will implement Computer Vision or Natural Language Processing models using Python and associated AI libraries.
Project choices will span domains including:
See below for project options that students could work on.
Project Example: Detecting Anti-Refugee Tweets using Sentiment Analysis
DATA SCIENCE PIPELINE FOR A REAL-WORLD PROBLEM
The project will help students understanding the data science pipeline for a real-world problem:
Data Cleaning and Visualization
Building the Initial AI Model
Interpreting Results and Improving the AI Model
Creating and Delivering Final Presentations and Results
Final Day: Students present their project to parents and guests on the final day. They complete a shareable digital poster and receive suggestions on taking the project forward.
AI for Social Good Projects
Projects will draw inspiration from topical examples such as the ones below.
AI for Eye Disease Diagnosis
Diabetic retinopathy is a dangerous eye condition that can lead to blindness if untreated. Detecting the disease typically requires a specialist eye doctor, and only 11 eye doctors exists for every 1 million people in India*. The use of computer vision based Artificial Intelligence models can automate this process allowing significantly more patients to be screened. We would work with deep learning models for detecting various health conditions.
*Source: New York Times
AI for Distracted Driver Detection
According to the National Highway Traffic Safety Administration (NHTSA) 3,450 people were killed in 2016 due to motor vehicle crashes involving distracted drivers. Innovations in Smart Car trends using AI allow distracted driver detection through a dashboard camera that can alert drivers in dangerous situations (e.g. when falling asleep). We would build deep learning models that allow us to accomplish such tasks.
AI for Identifying Crop Disease
Crop disease detection plays a vital role in ensuring farmer security. Unhealthy crops can lead to poor yields, and often leave farmers in a devastating predicament. Using computer vision on plant imagery, farmers can automatically predict whether a plant is diseased or not. This project would guide students to build a computer vision model to do so.
AI for Disaster Relief Improvement
Natural disasters like floods, earthquakes, etc. are times when aid needs to be provided urgently. Due to the volume of SMS messages, Tweets, Facebook posts, and newspaper articles, it can be a significant challenge to process these requests in a timely manner. The speed of processing can be the difference between life and death. AI models can be used to efficiently determine whether a message is actually a request for aid, and if so, can accurately classify the aid as food, clothing, water, etc. so that requests can be appropriately routed.