Save.AI

SAVE.AI is a platform that allows users to control cyberbullying on social platforms like Twitter, Facebook.

MODULES:

TWITTER

– In this module we are targeting the users who are tweeting malicious messages to the users so that they can harass them. We have trained our model using LSTM Recurrent Neural Network so that we can track their activities and if they surpass the safe user limit they would be retweeted automatically as a warning message.

FACEBOOK

– In this module we are targeting the users who are tagging users on explicit posts. First we would pass the image post through deep.ai and if that proves to be explicit and we would block that user. We would also check for image moderation for any nudity content in the images which are to be uploaded, it would be discarded immediately.

AI THERAPIST CHATBOT - EVA

Any person undergoing through depression can talk with Eva and she can effectively deal with issues regarding stress, loss and aggression.

She can solve simple emotions related problems which the user discusses with her. This healthy interaction with our therapist chatbot would pave way in helping the person overcome his anxieties and become confident.

The person can share his secrets with her which he is not comfortable sharing with his family and friends.

WEB APP: Interface to connect users and use safety features.

FACEBOOK Statistics:

-Username -Number of Posts -Profile url -Number of Bullies -Successful Detections

TWITTER Statistics:

-Username -Number of tweets -Number of bullied tweets -Safe tweets -Suspected Users

Dependencies

• TensorFlow

• Keras

• NumPy

• Flask

• Protobuf

• Jupyter notebook

• Dialog Flow

• Tweepy

• DeepAI