Millions of tonnes of plastic waste accumulate in landfills and oceans every year. One promising response is to engineer microbes to break the plastic down into useful chemical building blocks. However, teaching a bacterium to digest plastic efficiently demands fine-tuning not just one gene, but entire clusters of genes working in concert, like upgrading every machine on a factory assembly line rather than swapping out a single part.

A new platform developed by researchers from the National University of Singapore (NUS) could make that possible. Called Lytic Selection and Evolution (LySE), the system harnesses a modified bacteriophage — a virus that infects bacteria — to rapidly create and test many small genetic changes. It can improve long stretches of DNA (up to about 40,000 DNA letters), big enough to include most sets of genes needed for important chemical processes in cells.

A new machine-learning-based approach to mapping forests in high resolution and simulating their future growth, developed at Michigan State University and Virginia Tech, could help landowners figure out which trees to cut and which to leave standing to enhance pine profits.

“AI should be able to say ‘I’m Not Sure’ on its own.”

A new approach has been proposed to address the problem of “overconfidence”—one of the most critical risks of artificial intelligence (AI) in areas such as autonomous driving and medical diagnosis, where AI shows high confidence in incorrect predictions. A KAIST research team has developed a training method that enables AI to recognize situations involving unfamiliar or unseen knowledge, laying the foundation for reducing overconfidence and improving reliability.