Abstract 1– ML Enabled Food Journal (Approved)
In today's world, people are very health conscious and they want to remain fit. One of the main aspects of improving health is maintaining weight which is closely related to total calorie intake and burned by an individual.This problem offers a need for a tool for individuals to evaluate their daily calorie surplus or deficit.
Our solution leverages image classification to identify foods and log nutritional info into user’s journal.This simplifies calorie counting and provides a dataset to run analytics. Future improvements include pulling in health and activity data from wearables and smart scales. Predictive analytics could then be use to estimate weight loss and health based on real-time nutrition and fitness data. The application would leverage the cloud to score images using pre-trained DNNs, and store user’s data. A smartphone based front end application would be required.
With help of the solution, individual will not only know about their daily calorie consumption and calculate the calorie burned in the day but also get predicted weight loss or gain for a future date.
Abstract 2- Machine Learning backed fact checking and claim detection.
Using Natural Language Processing and Machine learning to classify the claim or quote as fact or fake. Leveraging statistical databases claims can be validated and scored to mark those as facts or fake. This can be used by users to do write better or most importantly to validate the information they are consuming.
Link to the notebook
http://ec2-52-32-59-130.us-west-2.compute.amazonaws.com:8888/notebooks/weigth-loss.ipynb
password : 1234