Today in Generative Media

Meta AI to get celeb voices; Fake AI podcasters; Will AI bust? States vs. election AI

News

• Exclusive: Meta's AI chatbot to start speaking in the voices of Judi Dench, John Cena, others, source says (Reuters)

• Fake AI “podcasters” are reviewing my book and it’s freaking me out (Ars Technica)

• Will A.I. Be a Bust? A Wall Street Skeptic Rings the Alarm. (New York Times)

• Half of U.S. states seek to crack down on AI in elections (Axios)

• NC governor candidate cries AI fabrication as defense for racist porn forum posts (Ars Technica)

• New Cloudflare Tools Let Sites Detect and Block AI Bots for Free (Wired)

• AI can generate recipes that can be deadly. Food bloggers are not happy (NPR)

• Civitai Gen-AI Makes Its Move (Forbes)

• ByteDance unveils 2 new video-generation AI models to narrow gap with OpenAI’s Sora (South China Morning Post)

Software

• PDF to Audio Converter This code can be used to convert PDFs into audio podcasts, lectures, summaries, and more (GitHub) Demo on HuggingFace.

• We've increased the size of our NVIDIA A100 fleet for paid users by around 2x, and for the last several days we've seen 100% success rate for users requesting A100s. (Google Colaboratory on X)

• Diffusers Outpaint now allows for infinite zoom-out with a resize input size + "use as input" button (X)

• Welcome to OpenMusic, a next-gen diffusion model designed to generate high-quality music audio from text descriptions! (HuggingFace)

Research

• V3: Viewing Volumetric Videos on Mobiles via Streamable 2D Dynamic Gaussians (SIGGRAPH Asia 2024, project page)

• PortraitGen Portrait Video Editing Empowered by Multimodal Generative Priors (SIGGRAPH Asia 2024, project page)

• MaskedMimic: Unified Physics-Based Character Control Through Masked Motion Inpainting (SIGGRAPH Asia 2024, project page)

• Colorful Diffuse Intrinsic Image Decomposition in the Wild (project page)

• MaterialFusion: Enhancing Inverse Rendering with Material Diffusion Priors (project page)

• Synergy and Synchrony in Couple Dances (project page)

• Are Pose Estimators Ready for the Open World? STAGE: Synthetic Data Generation Toolkit for Auditing 3D Human Pose Estimators (project page)

• Quality-aware Masked Diffusion Transformer for Enhanced Music Generation (project page). Code on GitHub. Demo on HuggingFace.

• SplatLoc: 3D Gaussian Splatting-based Visual Localization for Augmented Reality (project page)

• MGSO: Monocular Real-time Photometric SLAM with Efficient 3D Gaussian Splatting (arXiv)

• Dynamic 2D Gaussians: Geometrically accurate radiance fields for dynamic objects (arXiv)

Misc

• prompt: single celled organism seen under a microscope, microscopic world, strange bacteria creatures (X)

• AI will revolutionize creative work in e-commerce and fashion. (X)