AI Roundup 10/06 -> 10/13/2022

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

1. Microsoft integrates DALL-E 2 to its Microsoft 365 suite by introducing a Canva-like web app called Designer, along with Image Creator - a tool to generate images in Bing (TechCrunch)
   

   

2. Microsoft leverages OpenAI's Codex to uplevel its Power Automate and allows users to describe the flow they would like to build using natural language (TechCrunch)
   

   

3. Google open sources Crossmodal-3600 - a geographically-diverse set of 3600 images annotated with human-generated captions in 36 languages (Google AI)
   

   

4. Researchers from Meta AI and the University of Washington present "Self Ask" - a new method for improving language models' compositional reasoning (Paper)
   

   

5. A researcher from Peking University releases Stable-Dreamfusion - a PyTorch implementation of the text-to-3D model Dreamfusion (GitHub repo)
   

   

6. A voice synthesis company publishes an AI-generated podcast interview between Joe Rogan and Steve Jobs using their digitally cloned voices (Ars Technica)

7. Researchers from Tel Aviv University introduce Motion Diffusion Model (MDM)—a generative model that creates natural-looking human animations from textual inputs and existing animations (Marketpost)

   

8. Harmonai, a Stability AI-backed organization, releases Dance Diffusion - an algorithm and set of tools that can generate clips of music by training on hundreds of hours of existing songs (TechCrunch)

Last week’s AI Tidbits