DeepMoji, Artificial emotional intelligence with machine learning

Have you ever wanted to automatically detect the tone of a user in their messages? It’s definitely a tough ride! One word or one comma can change the meaning of a sentence.

Take an example, “This is shit” means Its bad and you don’t like it. “This is the shit” means It’s awesome and you love it! One word, “The” can change the meaning of the entire sentence. There are many slang words and edge cases to handle when dealing with artificial emotional intelligence.

This is when DeepMoji comes into the picture! DeepMoji is a pre-trained machine learning model which is open sourced and available for every one to use including companies.

DeepMoji demo

DeepMoji uses millions of texts on Twitter containing emojis for training a deep learning model that understands many nuances of how language is used to express emotions. For instance, it does well at capturing sarcasm and slang. It beats state-of-the-art algorithms across many benchmarks datasets.

DeepMoji can understand love, hate, sadness, even sarcasm!

Here’s a video explaining more about DeepMoji

DeepMoji is made by the team at MIT and they have open-sourced this aweome pre-trained model for Keras and PyTorch

Open source models

Keras pre-trainer model of DeepMoji –

PyTorch implementation of DeepMoji –

They call it as TorchMoji. Cool name!

Wrapping up

Check out their open-source pre-trained models and try building something cool using it! Don’t build Chitti the Robot with it 😂 If you did not understand what is Chitti the robot then I’m talking about a similar situation from iRobot movie where robots wage a war because of their emotional intelligence 🤖

Checkout more Dev resources from Coding is Love

3.5 2 votes
Article Rating
Notify of
Inline Feedbacks
View all comments