Facebook has taken steps to eradicate malicious content from the platform and developed a new system of identifying such videos and images and take care of them with the help of Artificial Intelligence (AI) and Machine Learning System (MLS).
Being one of the top social media sites, Facebook caters a lot of content that might be inappropriate to a lot of people online. Hence the company is continuously working towards combating this situation.
Facebook's recent blog post has described this system. Its named as Rosetta and takes the game of digital recognition a step ahead of optical character recognition (OCR). This system of mechanics works towards understanding the context of the content of the images along with the text that accompanies them.
Over a billion videos and images have been used to train this MLS. These images were taken from Instagram along with those of Facebook itself. Recognition takes place by the detection of a rectangular part with some text in most cases, which is caught up by the machine which then recognizes the text with the help of CNN (convolutional neural network).
Being one of the top social media sites, Facebook caters a lot of content that might be inappropriate to a lot of people online. Hence the company is continuously working towards combating this situation.
Facebook's recent blog post has described this system. Its named as Rosetta and takes the game of digital recognition a step ahead of optical character recognition (OCR). This system of mechanics works towards understanding the context of the content of the images along with the text that accompanies them.
Over a billion videos and images have been used to train this MLS. These images were taken from Instagram along with those of Facebook itself. Recognition takes place by the detection of a rectangular part with some text in most cases, which is caught up by the machine which then recognizes the text with the help of CNN (convolutional neural network).
Also Read: 4 in 10 Facebook Users Have Taken a Break from the Social Network in the past YearVarious products have already been integrated with Rosetta on Facebook as well as Instagram. This approach may be used to get better results in finding photos, and identifying hate speech in languages not only in English but Hindi and Arabic as well.
