Weekly AI news — October 23–29

CNN can match Transformers, IBM presents a new chip and much more

Photo by Possessed Photography on Unsplash

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…

Research

• Geographical erasure in language generation LLMs encode a vast amount of knowledge but it is not representative of all countries, Amazon shows how to mitigate this unbalanced
• Entangled Preferences: The History and Risks of Reinforcement Learning and Human Feedback. A deep dive into the history of RLHF, potential issues, and suggestions for new lines of research
• AgentTuning: Enabling Generalized Agent Abilities for LLMs. Open-source models are inferior to AI agents when you need them as efficient controllers for complex tasks. This paper highlights how to create efficient agent LLaMA-2 models
• The Foundation Model Transparency Index. Stanford’s new index rates the transparency of 10 foundation model companies and finds them lacking. The new index…