Other social networks also actively use ML

[tanjimajuha20]

"Artificial intelligence is a marketing term that usually refers to any technology based on machine learning. Machine learning is possible even without the use of neural networks. For example, to search for toxic comments, you can use the support vector method, which will allow you to identify negative statements after training. But the support vector method is a very outdated technology. Modern neural network architectures cope with this task much better. Most likely, they do not even use the "hyped" generative networks, but something simpler and more productive," says Alexander Krushinsky, Director of the Voice Digital Technologies Department of BSS LLC.

and neural network greece telegram technologies. For example, they are in VKontakte products. The social network has also implemented a development that recognizes offensive comments and hides them in a separate block. "The neural network operates on the basis of an ML model trained on a unique dataset consisting of tens of millions of comments, and can determine the context," said a representative of the VKontakte press service.

In addition, neural networks in social networks are used for sound processing (message decoding, noise reduction, automatic subtitle generation), translation, searching for objects in images, content recommendations, recommendations for adding as friends or candidates for consideration for a vacancy. Content moderation using AI is also important so that companies do not have problems with the law and do not violate the rules of the platform, noted Artem Valov.

However, the news feed can be formed not only with the help of AI-based technologies, Anton Frolov explained: "The feed can be built on the principle of thematic filtering. This is more like a linear process that does not relate to the sphere of artificial intelligence." Also, comments can be moderated only by specified words.

AI is most effective in applications that require processing large amounts of data to personalize content and improve user interactions, says Timur Shovgurov. In his opinion, algorithms in TikTok and YouTube have made great strides in creating unique content for each user, which has significantly increased engagement. Facebook* and Instagram* also use similar algorithms in feeds, and are also developing AI-based face recognition and content filtering technologies. Anton Frolov clarified that AI-based tools are most effective in large international and Russian platforms, as such companies have an extensive database necessary for training and also have the resources to implement AI technologies.

* Belongs to the Meta company, which is recognized as extremist and banned in the Russian Federati