On this page we give you the lastest news in AI. We cover new research, ethical dilemmas, business opportunities and everything else related to AI. If you want to receive this as a newsletter, you can sign up here.

: A survey of advances in vision-based vehicle re-identification3 weeks ago

Our take: This paper reviews how good surveillance has gotten lately. It is fascinating and opens the possibility to much more than just surveillance for security.

Facebook Research: Does Object Recognition Work for Everyone?3 weeks ago

Our take: This is a really good about the bias in AI - specifically, the bias in the data engineers. Why is that AI is better at Western cars, languages and animals than Eastern cars, languages and animals?

Huge Collaboration: Tackling Climate Change with Machine Learning3 weeks ago

Our take: We chose this article to be the last piece of content in the last newsletter before summer vacation. What a better way to end than this massive collection of how machine learning can help tackling climate change? Thanks to Alexander Nedergaard for sharing this paper. We still look forward to meeting you in person one day.

MIT: Training a single AI model can emit as much carbon as five cars in their lifetimes3 weeks ago

Our take: You might have read that mining blockchains carries a massive impact on the environment. This paper try to show a similar impact on the environment when training neural network. There are many flaws in the article. Firstly, the neural networks in question are massive networks that have already been trained once so that others do not need train it again. Secondly, the cloud vendors (Google, Amazon, Microsoft) have set up their data center to be carbon neutral. Thirdly, when you look at the research that we are often covering it seems that neural network are getting faster, better and quicker to train leading to less environmental impact.

⬅ Previous 1 of 19 pages. Next ➡