A number of neurals have recently been introduced, creating images by text description. The Meta Engineering Team has gone further and created the Make-A-Video algorithm, which is capable of creating small video clips by text description. At this stage, it is thus able to obtain low-quality video with artifacts, but in the future, the development can help to make significant progress in content creation with IE.
," says Meta*.
Director-General Meta* Mark Zuckerberg called the work on Make-A-Video, noting that making the video was much more difficult than the photo. The video generated by the present algorithm does not last more than five seconds and does not have sound escorts. The source notes that Meta* has yet to make the Make-A-Video neurize publicly available. This may mean that the videos published by the company could have been carefully selected to present the algorithm in a better light. It is expected that the quality of the video produced by the neuronet will improve in the future.
In Meta* it is believed that similar Mace-A-Video video tools can be invaluable to content creators and artists. However, as in the case of characterization algorithms, such tools can be dangerous. This is because such algorithms can be used for disinformation, propaganda, etc. You can get to know a large number of videos here.
* Listed on the list of voluntary associations and religious organizations in respect of which the court has taken a legally enforceable decision to abolish or prohibit activities on the grounds provided for in Federal Act No. 114-FZ of 25 July 2002 on countering extremist activities.