How Twitter’s Algorithm Detect Important Actions in a Stream?

Twitter has been working on technology that will recognize the activities in a live video. As someone streams an item, Twitter could allow you to find it right away. Meerkat and Periscope tend not to tag as well as they could but thanks to the AI team at Twitter – Cortex – has an algorithm that detects the action in a stream and as content regularly changes on live videos, it is a great thing that they can now be tagged.

It is impressive progress that has been made recently as identifying video content is harder than identifying items in pictures. It takes a lot of skill to do it right away A computer has been built specifically or this and according to Peter Brusilorvsky from the University of Pittsburgh videos are not usually easy to skim, and processing power is important.

READ ALSO: China Will build its First Maritime Nuclear Power Platform

The next aim for the Cortex team is to develop a system that will filter content and share it based on the viewers previous viewing history. Tests are being carried out on Periscope, which is a Twitter mobile phone app.

Videos can be tagged in many ways and as a result, it can be hard to find good clips if they are not tagged as you would expect. As yet there is no advertising on Periscope, but that may have to change, and it will be easy to add the correct form off advertising once the content has been correctly catalogued.

Professor Ben Edelman of Harvard’s Berkman Centre believes it will be crucial when it comes to filtering out pornography.

Related Coverage

• Researchers Invent Smartwatch Algorithm That Records Your Every Move
• Think Before Tag Someone in a Facebook Photo, It Can Cause Some Jail Time
• Drones Get Clever by Stealing Data from PC’s through its Blinking Lights
• Are Twinkling Little Enzyme’s the Answer to Curing Cancer?
• UCLA Released a Powerful Image Detection Algorithm For Everybody’s Use