WEEKLY AI NEWS: RESEARCH, NEWS, RESOURCES, AND PERSPECTIVES
LLMs for biology, brain-like neural networks, OpenAI new board, and much more
The most interesting news, repository, articles, and resources of the week
Check and star this repository where the news will be collected and indexed:
GitHub — SalvatoreRa/ML-news-of-the-week: A collection of the the best ML news every week…
A collection of the the best ML news every week (research, news, resources) — GitHub — SalvatoreRa/ML-news-of-the-week…
You will find the news first in GitHub. Single posts are also collected here:
- SegVol: Universal and Interactive Volumetric Medical Image Segmentation. Clinical analysis has entered a new era with the release of SegVol, a universal model for medical picture segmentation. SegVol is highly proficient at segmenting a wide range of anatomical categories, having been trained on a large set of CT images. official code.
- Visual In-Context Prompting. This novel strategy supports a variety of cues and environments, significantly improving performance in segmentation tasks and demonstrating outstanding outcomes in open-ended challenges. official code.
- Starling-7B: Increasing LLM Helpfulness & Harmlessness with RLAIF. Researchers at Berkeley used synthetic preference data to train a brand-new, cutting-edge 7B parameter model. This blog discusses the unique difficulties in training reward models (such as how an example’s score might change depending on where it is in the list) and how they overcome them. Both the training reward model and the generated model are made public.
- Segmentation-Based Parametric Painting. A novel method has been devised by researchers to convert pictures into paintings that emulate human characteristics and aesthetics.