AI Roundup 09/08 -> 09/15/2022

Last week's most interesting AI news in <2 minutes

1. Amazon released a new multilingual language model, AlexaTM 20B, outperforming GPT-3 in translation and text summarization tasks (Amazon’s Science blog)

   

2. Google showcases AudioLM, a generative AI model for audio that can generate speech and piano music continuations given short prompts (Google Research)

3. Microsoft's recently released X-CLIP, a framework for video-language understanding, is now available on Hugging Face (Hugging Face)

   

4. Runway teases AI-powered text-to-video editing using written prompts (Ars Technica)

   

5. Meta will spin out PyTorch to an outside operator under the broader Linux Foundation consortium (Meta’s blog)

6. Meet Phraser: a new tool for writing better prompts (Marktechpost)

   

7. Google AI present MaxViT - a Multi-Axis Vision Transformer that can better adapt to high-resolution, dense prediction tasks, and different input sizes (Google AI blog)

8. Researchers from the Chinese startup Meituan released YOLOv6 - an object detection framework displaying better accuracy with lower latency (Github)

Last week’s AI Tidbits