Disturbance Detection in Messaging Apps Using Machine Learning

A Fully Deployable React-Native mobile app that seeks to classify incoming messages in messaging apps into important or disturbing categories. using a Multi-Modal Machine Learning Architecture to achieve Text classification, Image classification and YouTube Video Link classification.

Tech Stacks Used:

1. React Native - App Dev
2. GraphQL - API Querying
3. AWS Amplify - Backend
4. Google BERT Algorithm - Text Classification
5. ResNet 34 - Image Classification

Code Execution Instructions:

1. The cloned repo can be opened on Visual Studio Code, Atom or any standard IDE.
2. Make sure Expo is installed in the target device. Please follow the Support Document for further assistance in this regard.
3. Install the Expo App (available on both iOS as well as Android) in your mobile device.
4. Open terminal. Navigate to cloned repo. Type in the command expo start.
5. It should immediately launch the app server in a localhost of your default browser.
6. Open the Expo App keeping your mobile device connected to the same network as your PC.
7. Deep Chat will show up in the list of running apps.
8. Click on it to launch the app.

Alternatively, you can...

9. Visit expo.dev and create an account.
10. Open terminal. Navigate to cloned repo. Type in the command expo build:android or expo build:ios depending on your mobile platform.
11. This will start a build of the project and publish it in the site mentioned in Step 9.
12. You can download this file and install it in your phone just like any other app.