Focus on filtering real-time content
Facebook is focusing on effectively trying to filter content including live video streams in real-time. The company would like to take down offensive videos of its features such as live, that contain murder, suicide, violence and hate speech. They want to take down content as quickly as possible as the events happen. And for this, they are building the custom chips that will help them with their main focus right now. Creating their own chips would allow Facebook to spot this kind of content faster and more efficiently. These chips along with Machine Learning (ML) processes will help with such issues. AI is helping Facebook a lot to tackle such issues to the point where its efforts can detect suicidal tendencies. Still, it’s not sophisticated and efficient enough to handle a lot of the pressing issues the network faces.
Machine Learning (ML)
Facebook plans on designing its own machine learning (ML) processors to make the chips for filtering content. This is not new for Facebook as it has already produced hardware such as motherboards and communications chips for data centers and servers. It was reported last month by Bloomberg, the company was looking to put together a team to work on semiconductors. But Facebook needs to produce the computer chips that are energy efficient, which conventional methods using human moderators aren’t. By producing its own chips, it can not only remove content quickly and efficiently but also integrate them in various other hardware products.
Software Companies Embrace Machine Learning (ML) Hardware
We are seeing more and more software companies making their own chips and embracing machine learning processes. Tech giants such as Google, Apple, Amazon and Microsoft and even smartphone makers are focusing on ML hardware. The hardware helps in improving features such as photo quality and battery life. The trend does not seem to show any signs of slowing down so we can expect a bright future in the world of technology.