AI Roundup 09/22 -> 09/29/2022

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

1. Meta introduces Make-A-Video: An AI system that generates videos from text (Meta AI)

   

2. OpenAI opened access to its DALL-E 2 playground by removing its waitlist (OpenAI)

3. Google releases DreamBooth: Fine Tuning Text-to-Image Diffusion Models for Subject-Driven Generation (Marketechpost)

   

4. DeepMind published a research paper outlining Sparrow, a new chatbot that reduces the risk of unsafe and inappropriate answers (DeepMind's blog)

   

5. Hugging Face and ServiceNow launch BigCode to open source code-generating AI like AlphaCode and Codex (TechCrunch)

6. Meta AI releases EditEval, an instruction-based evaluation suite for automatic evaluation of editing capabilities (EvalAI)

7. A researcher used Stable Diffusion to compress images to 1% of their original size (Ars Technica)

   

8. Private medical records were found in the web-scraped LAION 5B images dataset (Ars Technica)

9. According to Microsoft’s chief science officer, Interactive & Compositional deepfakes are two new deepfake threats (VentureBeat)

10. NVIDIA Research builds AI model (GET3D) to populate virtual worlds with 3D objects (Nvidia AI)

Last week’s AI Tidbits