Weekly AI Pulse #34, AI Beauty Contest, Microsoft New Video Generator and More!

Your Best Friend to Catch Up AI News

Welcome to your best friend to catch up on AI News! As you know from our previous ones, in Weekly AI Pulse, you don’t need to read the article; you can also listen.

For a while, we have been trying to add one more feature. With that, you can also watch our news too, like this one;

Microsoft's Vasa-1

Microsoft researchers have created VASA-1, an innovative AI tool that produces realistic videos from still images. While the technology impressively mimics human speech and movements, it struggles with rendering teeth accurately.

These often appear cartoonish, which detracts from the otherwise lifelike quality of the videos.

Highlights from the Research:

• VASA-1: Takes a still photo and an audio file to create a video with synchronized facial movements.

• Quality and Speed: Generates high-quality 512x512 videos at up to 40 FPS with minimal delay, ideal for real-time applications like translation services.

• Challenges with Teeth: Teeth in the videos appear overly animated, especially noticeable when slowed down.

Microsoft is cautious about releasing this tool publicly due to potential misuse, particularly with concerns about deepfakes as the U.S. presidential election approaches.

Listen

By doing so, you’ll have the option of watching AI news.