2-Min AI Newsletter #14

???? ?? Latest AI/ML Research Highlights?

???Latest Computer Vision Research at Nanyang Technological University Introduces VToonify Framework for Style Controllable High-Resolution Video Toonification

Researchers suggest a brand-new hybrid framework for video toonification called VToonify. To get beyond the fixed crop issue, they must first examine the translation equivariance in StyleGAN. They use the StyleGAN architecture for high-resolution style transfer, as shown below. However, they modify StyleGAN by removing its fixed-sized input feature and low-resolution layers to create a fully convolutional encoder-generator architecture

???Meta AI Introduces ‘Make-A-Video’: An Artificial Intelligence System That Generates Videos From Text

Make-A-Video is a new AI technology that enables individuals to convert text suggestions into brief, high-quality video snippets. Make-A-Video advances recent developments in Meta AI’s research on generative technologies.?A multimodal generative AI technique allows users more control over the AI-generated material they create.

???The White House just unveiled a new AI Bill of Rights

The White House on Tuesday released its?long-awaited?“bill of rights” on artificial intelligence, laying out the Biden administration’s vision for how companies and government agencies can safeguard people from misuse or abuse at the hands of the technology.?

???Meta launches AI software tools to ease switching between Nvidia, AMD chips

Facebook parent Meta Platforms Inc?(META.O)said on Monday?it has launched a new set of free software tools for artificial intelligence applications that could make it easier for developers to switch back and forth between different underlying chips. Meta's new open-source AI platform is based on an open-source machine learning framework called PyTorch, and can help code run up to 12 times faster on Nvidia Corp's?(NVDA.O)?flagship A100 chip or up to four times faster on Advanced Micro Devices Inc's?(AMD.O)?MI250 chip, it said.

???Researchers at Columbia University Propose the Unbounded Depth Neural Network (UDN): An Infinitely Deep Probabilistic Model that Enables Self-Adaption to the Training Data Without an Upper Limit

Researchers propose the new unbounded depth neural network (UDN), an infinitely deep probabilistic model that enables self-adaptation to the training data without an upper limit, saving deep learning researchers and practitioners from difficult decisions regarding the appropriate complexity for their DNN architectures in the new paper Variational Inference for Infinitely Deep Neural Networks.

??Top Artificial Intelligence (AI) Chip Companies Leading The Way In 2022

Future economic growth could be significantly impacted by AI chips, which will unavoidably be used in robotics, smart homes where electronic items are becoming more intelligent, cars that are becoming more autonomous, and many other technologies. As the name suggests, AI chips are a new generation of microprocessors created primarily to handle tasks requiring artificial intelligence more quickly and efficiently.

????Cool AI Papers From Last Week

???WaveFit: An Iterative and Non-autoregressive Neural Vocoder..

???Improving Sample Quality of Diffusion Model Using Self-Attention Guidance

? IntrinsicNeRF: Learning Intrinsic Neural Radiance Fields for Editable Novel ..

???PromptKG: A Prompt Learning Framework for Knowledge Graph Representation.

???TabDDPM: Modelling Tabular Data with Diffusion Models