Narrativa launches on Hugging Face with hate speech detection model

We are proud to announce that as of today, Narrativa has launched on the Hugging Face platform with an artificial intelligence model that detects hate speech on Twitter. A problem that is impacting more and more users, over half of Americans (53%) experienced harassment and/or hate speech in 2018 per USA Today. Narrativa is offering new solutions to this growing problem.

What is Hugging Face?

The US platform Hugging Face was born in 2016. Its objective is to democratize artificial intelligence and make science something universal and accessible to all, an idea that Narrativa shares and that has already been put into practice with its coronavirus monitoring project.

Our NLP engineer, Manuel Romero, is the biggest contributor ever to the platform (he has already created more than 200 models!). In his profile you can see some, such as the summarization model in Spanish.

Artificial intelligence to detect hate speech

Hate speech is widespread on social media. It is common to find racist, homophobic and sexist messages that violate the rules and regulations of social media networks, but is often impossible to detect due to the large volume of posts made daily.

At Narrativa we have created a model capable of detecting these messages on Twitter. These are some of the applications that this model could have in real life:

• Internet forums: It is common for discussion forums to use very conversational speech that includes hate speech. It goes unpunished if there is no moderation.
• Comments section in the media: While the media do have moderators to ensure that the rules and regulations are not violated, it is still difficult to read all the comments that are written. With this system, the moderator could be alerted to messages that violate the norm and thus facilitate their work.

Our model is very advanced, as we have one of the most powerful metrics regarding datasets on the market.

Where can I download the model?

Click here to access Narrativa on Hugging Face.

To access the hate speech detection model, click here.

Social role of artificial intelligence

We often talk about artificial intelligence having the ability to change the status quo. While there are many negative messages on the internet, AI is a tool that can contribute to ending hate speech and making a more peaceful society.