Suggested readings week 40/2022
AI-generated algorithms
So here we are, in an anticipated and (by someone) feared step forward in the direction of algorithms that evolve themselves.
In our paper, published today in Nature, we introduce AlphaTensor, the first artificial intelligence (AI) system for discovering novel, efficient, and provably correct algorithms for fundamental tasks such as matrix multiplication.
An algorithm based on Artificial Intelligence (Machine Learning) generates novel algorithms that might help preparing faster Artificial Intelligence solutions. This is not totally yet true because apparently the order of magnitude of the problem solved here is out of the scope of existing Artificial Intelligence complexity.
Nonetheless the road is clear, along with ethical and practical implications. Not to mention the observability problem: how to be sure that a generated solution is always correct?
https://www.deepmind.com/blog/discovering-novel-algorithms-with-alphatensor?utm_source=pocket_mylist
Espresso
A scientific approach to problem understanding and solving can be learned with a bit of fun from this deeply rigorous description of how Espresso (yes, coffee) is prepared. Building a mindset: learn and improve.
https://www.cell.com/matter/pdf/S2590-2385(19)30410-2.pdf?utm_source=pocket_mylist
Birds and DNA
This is a fascinating story of DNA, birds and wildlife observation published by one of the most fascinating newspapers.
Economic consequences of fighting climate change
Disclaimer: this relates to US economic ecosystem, but might well scale to Europe very soon.