In the Arizona desert, a pair of robots methodically trundles back-and-forth across the craggy earth. Bulky, angular and slow, they're not terribly impressive to watch. But U.S. Army leaders see these robots as a vision of the future: part of a new pipeline to put better, more reliable technology into the hands of soldiers faster than ever before. A year earlier, at the first-ever Project Convergence, held in 2020 at Yuma Proving Ground, users had to tell the robot to go from point A to point B to point C to conduct a reconnaissance mission. For the 2021 event, users simply gave the robots a designated area for the same task, and the system turned to artificial intelligence to determine the best path. The robots demonstrated how they could keep soldiers out of harm's way, allow for sensors in new positions that were previously impractical and present new data to commanders.

A corkscrew-shaped microrobot inspired by the tails of bacteria like E. coli can swim through blood vessels and help unblock clots. Li Zhang at the Chinese University of Hong Kong and his colleagues inserted the robot into a synthetic vein filled with pig's blood, and found it made blood clot-busting drugs work nearly five times better than the drug by itself. "The helical structure is just like a propeller, so [the robot] can deliver the cargo from point A to …

Processing raw sensory inputs is crucial for applying deep RL algorithms to real-world problems. For example, autonomous vehicles must make decisions about how to drive safely given information flowing from cameras, radar, and microphones about the conditions of the road, traffic signals, and other cars and pedestrians. However, direct "end-to-end" RL that maps sensor data to actions (Figure 1, left) can be very difficult because the inputs are high-dimensional, noisy, and contain redundant information. Instead, the challenge is often broken down into two problems (Figure 1, right): (1) extract a representation of the sensory inputs that retains only the relevant information, and (2) perform RL with these representations of the inputs as the system state. Representation learning can extract compact representations of states for RL.

If you have ever dreamed of earning money from a stellar music career but were concerned you had little talent, don't let that put you off - a man called Alex Mitchell might be able to help. Mr Mitchell is the founder and boss of a website and app called Boomy, which helps its users create their own songs using artificial intelligence (AI) software that does most of the heavy lifting. You choose from a number of genres, click on "create song", and the AI will compose one for you in less than 30 seconds. It swiftly picks the track's key, chords and melody. You can do things such as add or strip-out instruments, change the tempo, adjust the volumes, add echoes, make everything sound brighter or softer, and lay down some vocals.

AI deployment in the enterprise skyrocketed as the pandemic accelerated organizations' digital transformation plans. Eighty-six percent of decision-makers told PricewaterhouseCoopers in a recent survey that AI is becoming a "mainstream technology" at their organization. A separate report by The AI Journal finds that most executives anticipate that AI will make business processes more efficient and help to create new business models and products. The emergence of "no-code" AI development platforms is fueling adoption in part. Designed to abstract away the programming typically required to create AI systems, no-code tools enable non-experts to develop machine learning models that can be used to predict inventory demand or extract text from business documents, for example.

It's 2022, and developments in the AI industry are off to a slow -- but nonetheless eventful -- start. While the spread of the Omicron variant put a damper on in-person conferences, enterprises aren't letting the pandemic disrupt the course of technological progress. John Deere previewed a tractor that uses AI to find a way to a field on its own and plow the soil without instructions. As Wired's Will Knight point outs, it and -- self-driving tractors like it -- could help to address the growing labor shortage in agriculture; employment of agriculture workers is expected to increase just 1% from 2019 to 2029. But they also raise questions about vendor lock-in and the role of human farmers alongside robots.

For years, automakers have told a specific story about how self-driving cars would arrive in the world. They would be shared and electric, fleets of ride-hail vehicles shuttling passengers like fancy taxis. General Motors and Lyft signed an agreement to pull it off back in 2016; Ford promised its robotaxis would debut by last year; Daimler said it would work with Uber to deploy fleets of Mercedes-Benzes. The logic was financial: Autonomous vehicle technology would be so expensive to develop that carmakers wouldn't be able to offer it to most drivers at prices they could afford. This vision carried profound implications: If city dwellers could depend on fleets of shared robotaxis for long trips, they could abandon the personal car altogether.

Hollywood's worst-case scenario involving artificial intelligence (AI) is familiar as a blockbuster sci-fi film: Machines acquire humanlike intelligence, achieving sentience, and inevitably turn into evil overlords that attempt to destroy the human race. This narrative capitalizes on our innate fear of technology, a reflection of the profound change that often accompanies new technological developments. However, as Malcolm Murdock, machine-learning engineer and author of the 2019 novel The Quantum Price, puts it, "AI doesn't have to be sentient to kill us all. There are plenty of other scenarios that will wipe us out before sentient AI becomes a problem." "We are entering dangerous and uncharted territory with the rise of surveillance and tracking through data, and we have almost no understanding of the potential implications."

If you have been considering getting into machine learning, now is the time to start. Demand for machine learning engineers (already in short supply) is high and only expected to grow as the complexity of, and access to, machine learning increases. Machine learning has changed rapidly over the past few years. Within the tech world, the bounds of machine learning are constantly being pushed. The complexity of machine learning models and systems engineering has increased as more applications demand real-time or near real-time inferences.