AI News: DALL-E generates pixels from the text. Now meet its cousin, VALL-E; Meet Med-PaLM- a Large Language Model....

This newsletter brings the latest developments in machine learning and artificial intelligence.

Stanford: Researchers at Stanford have developed an Artificial Intelligence (AI) Model, SUMMON, that can generate Multi-Object Scenes from a Sequence of Human Interaction.

Google Research: This research work demonstrates dramatic improvements in the generality and performance of learned optimizers, by scaling up meta-training compute and dataset size, and by making architectural improvements. The resulting learned optimizer, VeLO, has no hyperparameters, yet outperforms heavily hyperparameter-tuned baselines across more than 80 diverse optimization tasks, including large published machine learning models.

Zhejiang University/Alibaba: BoxInstSeg is a toolbox that aims to provide state-of-the-art box-supervised instance segmentation algorithms. It is built on top of mmdetection. The main branch works with Pytorch 1.6+ or higher (we recommend Pytorch 1.9.0)

Meta AI: Shares more details about the Variance Reduction System — a new offline RL framework developed by Meta to help ensure a more equitable distribution of ads on our services

Microsoft AI: Microsoft team introduces Federated Learning Utilities and Tools for Experimentation (FLUTE) as a framework for running large-scale offline federated learning simulations.

CVF: All papers of WACV 2023 are now available in - Open Access Repository provided by the Computer Vision Foundation.

Stable Diffusion Application: This New API Makes It Easy And Cheap For Developers To Build Machine Learning ML-Powered Apps Using Stable Diffusion.

Microsoft: DALL-E generates pixels from the text. Now meet its cousin, VALL-E, which generates audio from text. VALL-E Can Turn Text into Speech Using Three-Second.

Stanford: A key issue with LLMs is understanding how certain the model is about an answer. This paper demonstrates a proof-of-concept of a large language model conducting corporate lobbying-related activities.

Google/Deepmind: Med-PaLM is a Large Language Model that is Supporting the Medical Domain in Providing Safe and Helpful Answers.

Meta/KAIST/NYU: Convolutional networks strike back, again. The fully convolutional ConvNeXt v2 extends the successful ConvNeXt architecture by adding self-supervised learning capabilities. What's new?

💬 What do you think

🤔 Interesting Tweet

Sponsor our Newsletter | Join our 13k+ ML Subreddit