Research in the field of machine learning and AI, now a key technology in practically every industry and company, is far too voluminous for anyone to read it all. This column, Perceptron, aims to collect some of the most relevant recent discoveries and papers -- particularly in, but not limited to, artificial intelligence -- and explain why they matter. An "earable" that uses sonar to read facial expressions was among the projects that caught our eyes over these past few weeks. So did ProcTHOR, a framework from the Allen Institute for AI (AI2) that procedurally generates environments that can be used to train real-world robots. Among the other highlights, Meta created an AI system that can predict a protein's structure given a single amino acid sequence.

The days of having to remember to your ID with you could soon be a thing of the past. The famous tourist town of Wuzhen, China, is now using face recognition technology to act as its entry pass through the gates of the attraction. The system uses cameras to spot people as they approach the entry, and checks these against a database of registered visitors within a few seconds. The facial recognition technology is thought to be up to 99.77 per cent accurate and able to distinguish people better than a human. Chinese web firm Baidu's system is based on neural networks, which can process huge amounts of data – more than one billion faces – with 99.8 per cent accuracy.