Hello Nature readers, would you like to get this Briefing in your inbox free every day? Sign up here
Bees: “feed me or I’ll cut you”
When pollen is scarce, bumblebees have their way of extorting more from plants. The insects bite into leaves with their mandibles and proboscises to induce flowering up to one month earlier than normal. Chemical ecologists spotted the unusual behaviour in Bombus terrestris during an unrelated laboratory experiment. Artificial cuts were not as effective, suggesting that some chemicals in the insects’ saliva could play a part. “This is one of those really rare studies that observes a natural phenomenon that hadn’t been documented before,” says ecologist John Mola.
Scientific American | 4 min read
Dashed hopes for African AI conference
An April meeting planned to be held in Addis Ababa, Ethiopia, would have been the first major artificial-intelligence (AI) conference in an African country — until coronavirus forced it to go virtual. The location of the annual International Conference on Learning Representations would have made it more accessible to researchers who cannot readily get visas to Western countries, and given Ethiopia a powerful opportunity to boost its research environment. “Africans and the institutions in Africa missed a chance to make connections and secure important collaborations,” says computer scientist Meareg Hailemariam.
The proportion of 200 million tweets about the coronavirus that were probably from fake accounts designed to sow disinformation, according to an as-yet-unpublished analysis. (NPR | 4 min read)
Features & opinion
Symbolic maths yields to neural networks
Deep learning lets computers work out statistical patterns in massive amounts of data. It is a more brute-force approach than symbolic AI, in which programmers encode explicit rules in their algorithms. Now, two computer scientists have shown that deep learning can handle mathematical symbols, too. Borrowing techniques from automated translation, they taught their neural networks to solve mathematical problems, such as integration. “Mathematicians will in general be very impressed if these techniques allow them to solve problems that people could not solve before,” says mathematician Anders Hansen. Some hope that a similar strategy could enable computers to find their own mathematical proofs.
Source: arXiv preprint
Synthetic eye that ‘sees’ like a human
Electronic engineer Zhiyong Fan and his colleagues have built a biomimetic eye with a hemispherical retina made of perovskite nanowires. The shape gives the eye better image-sensing characteristics than a flat light sensor, and makes it appear more human. Electronic engineer Zhiyong Fan tells the Nature Podcast how he went from admiring robots on Star Trek to taking a step closer to making them a reality.
Nature Podcast | 22 min listen
Get the expert analysis from electronic engineer Hongrui Jiang in the Nature News & Views article.
Subscribe to the Nature Podcast on iTunes, Google Podcasts or Spotify.
Where I work