Facebook created over 10,000 ‘hate memes’ to help researchers learn how to spot hate speech (FB) [ARTICLE]



[ad_1]

  • They will be able to use this dataset to learn how to identify hate speech online to better protect themselves against it.
  • The dataset contains pieces of “multimodal content” content where its meaning is derived from multiple different elements, which may be difficult for AI to understand.
  • Facebook also released its latest community app report on Tuesday.
  • Facebook has created over 10,000 racist, sexist and hate memes and it’s for a good cause.

On Tuesday, Facebook announced that it has created a dataset of more than 10,000 “hate memes” that will be available to selected researchers working to tackle hate speech online. The database was announced alongside the company’s latest community compliance report, a report detailing the volume of harmful content the social media giant detects and removes from its platform, from hate speech to illegal material. .

Modern content moderation relies heavily on advanced machine learning and artificial intelligence, but such technologies generally need to be trained by displaying numerous examples of a particular type of content before they can reliably learn to recognize it. The “hate meme” dataset created by Facebook is intended to provide a readily available body of data for researchers who can then analyze it to create technology that can better detect it in the future. Facebook also rebuilt the memes using licensed Getty images to avoid copyright issues.

The dataset includes racist, sexist and incitement to violence material, Facebook said in a blog post: “Our examples also cover a wide variety of protected categories (such as religion, gender and sexual orientation) and types of attacks (such as inciting to violence or portraying types of people as criminals or terrorists. The distribution in the dataset reflects the real-world distribution found in the original examples. “

Memes are specific examples of what is called “multimodal content” content that derives its full meaning by taking into account different elements (eg text, images) at the same time. A meme can have a non-offensive title and a generic photo, but once combined in a certain way they become insulting or hateful.

In a very slight example that Facebook shared, a photo of an empty desert is captioned “See how many people love you.” Any of the elements taken in isolation would be harmless, but once combined, they become insulting.

Facebook says that given the sensitivity of the dataset, it will only be available to researchers who agree to the terms of use on how they use, share, and store it. It will not be available for the general public to download.

Facebook is also launching a contest for researchers, the Hateful Memes Challenge, with a prize of $ 100,000 to encourage them to develop AI models using the dataset.

LOOK NOW: Why is thoroughbred horse semen the most expensive liquid in the world?

See also:

SEE ALSO: Facebook’s fight against erroneous information about the coronavirus could increase pressure on the company to be more aggressive in eliminating other falsehoods spreading across the social network

[ad_2]