Like the development of a car model (say, electrical cars), developing AI models is a costly and time-consuming process. The lifecycle of an AI model can be divided into two phases: training and deployment. The training phase includes data collection and pre-processing, model selection (e.g., architecture search and design), hyperparameter tuning, model parameter optimization, and validation. AI model training can be quite expensive, especially when it comes to the training of foundation models4 that require pre-training on large-scale datasets with neural networks consisting of a gigantic size of trainable parameters. Take the Generative Pre-trained Transformer 3 (GPT-3)5 as an example, which is one of the largest languages models ever trained to date.

Researchers at The University of Texas at Dallas have used automated machine learning in a new way to forecast state violence in Africa, and they expect the technology to have even wider predictive applications. Dr. Vito D'Orazio, associate professor of political science in the School of Economic, Political and Policy Sciences, and his team created the dynamic forecasting model as part of a competition sponsored by the Violence Early-Warning System (ViEWS) project at Uppsala University's Department of Peace and Conflict Research. The research was subsequently published online Jan. 15 in the journal International Interactions. The ViEWS contest challenged competitors to forecast -- up to six months out -- the change in the number of fatalities in a country or region stemming from state-based violence, which is armed conflict in which at least one party is a government. Forecasts from the UT Dallas team were so accurate on the subnational level -- consisting of randomly gridded map areas that don't take countries' borders into account -- that they won that part of the competition for predictive accuracy and split the win for originality.

"The biomarker field is blessed with a plethora of imaging and molecular-based data, and at the same time, plagued with so much data that no one individual can comprehend it all," explained Guest Editor Karin Rodland, PhD, Pacific Northwest National Laboratory, Richland; and Oregon Health and Science University, Portland, OR, USA. "AI offers a solution to that problem, and it has the potential to uncover novel interactions that more accurately reflect the biology of cancer and other diseases." Promising applications of AI, DL, and ML presented in this issue include identifying early-stage cancers, inferring the site of the specific cancer, aiding in the assignment of appropriate therapeutic options for each patient, characterizing the tumor microenvironment, and predicting the response to immunotherapy. A comprehensive overview of the literature regarding the use of AI approaches to identify biomarkers for ovarian and pancreatic cancer illustrates underlying principles and looks at the gaps and challenges that face the field as a whole. Ovarian and pancreatic cancers are rare, but lethal because they lack early symptoms and detection.

Autonomous driving regulators in the US have finally made it official: fully self-driving transportation in the country will no longer have to equip human driving controls in order to meet traditional road safety standards. The momentous rule change means that for the first time, automated vehicle makers will no longer have to include manual driving controls such as brake pedals and steering wheels for vehicles designated fully self-driving, representing a mammoth shift in safety expectations and regulatory oversight of the autonomous driving market. According to the National Highway Traffic Safety Administration (NHTSA), the previous standard that all road vehicles had been held to, had actually been created in an era where fully autonomous driving had not even been conceived of -- essentially, the rules presumed that a vehicle "will always have a driver's seat, a steering wheel and accompanying steering column, or just one front outboard passenger seating position." These regulations have obviously become passe in an era where fully autonomous driving is now a distinct possibility, but even though the rules have been revised, lingering questions concerning road safety still remain. Last month, General Motors' self-driving subsidiary Cruise petitioned the NHTSA to approve and fully deploy its fully-autonomous Origin into full commercial operations -- a problem at the time since Origin did not come with human-centered features, like a steering wheel or a sun visor.

A meteorite has been successfully recovered using drones and a machine-learning algorithm for the first time. The development marks a step forward in improving the process of searching Earth's surface for extraterrestrial rocks. Typically, when a meteorite falls to Earth, it takes a team of five or more people to visually search the ground around the fall zone to find it. While the process can be successful, it is costly and time-consuming.

Few fields have been more filled with hype and bravado than artificial intelligence. It has flitted from fad to fad decade by decade, always promising the moon, and only occasionally delivering. "Let me start by saying a few things that seem obvious," Geoffrey Hinton, "Godfather" of deep learning, and one of the most celebrated scientists of our time, told a leading AI conference in Toronto in 2016. "If you work as a radiologist, you're like the coyote that's already over the edge of the cliff but hasn't looked down." Deep learning is so well-suited to reading images from MRIs and CT scans, he reasoned, that people should "stop training radiologists now" and that it's "just completely obvious within five years deep learning is going to do better."

Someone at a robot company once told me a story about one of its bomb disposal machines. The soldiers who had been using the robot in Afghanistan were dismayed after it returned from repairs. They said that the robot's shiny new parts and casing--lacking the bullet holes and blast scars they knew--made it seem as if the machine itself had, in a sense, died. It might seem odd, grieving a robot. But for anyone who's seen After Yang, the beautiful and strange new movie by the South Korean filmmaker Kogonada, it won't.

On March 1, Chechnya's leader Ramzan Kadyrov posted a short video on Telegram, in which a cheery bearded soldier stood before a line of tanks clanking down a road under an overcast sky. In an accompanying post, Kadyrov assured Ukrainians that the Russian army doesn't hurt civilians and that Vladimir Putin wants their country to determine its own fate. In France, the CEO of a law enforcement and military training company called Tactical Systems took a screenshot of the soldier's face and got to work. Within about an hour, using face recognition services available to anyone online, he identified that the soldier was likely Hussein Mezhidov, a Chechen commander close to Kadyrov involved in Russia's assault on Ukraine, and found his Instagram account. "Just having access to a computer and internet you can basically be like an intelligence agency from a film," says the CEO, who asked to be identified as YC to avoid potential repercussions for his sleuthing.

Machine learning techniques are providing new tools that could help archaeologists understand the past -- particularly when it comes to deciphering ancient texts. The latest example is an AI model created by Alphabet-subsidiary DeepMind that helps not only restore text that is missing from ancient Greek inscriptions but offers suggestions for when the text was written (within a 30-year period) and its possible geographic origins. "Inscriptions are really important because they are direct sources of evidence ... written directly by ancient people themselves," Thea Sommerschield, a historian and machine learning expert who helped created the model, told journalists in a press briefing. Due to their age, these texts are often damaged, making restoration a rewarding challenge. And because they are often inscribed on inorganic material like stone or metal, it means methods like radiocarbon dating can't be used to find out when they were written. "To solve these tasks, epigraphers look for textual and contextual parallels in similar inscriptions," said Sommerschield, who was co-lead on the work alongside DeepMind staff research scientist Yannis Assael.

Timnit Gebru, the founder of the Distributed Artificial Intelligence Research Institute, called for scholars to employ more ethical approaches in artificial intelligence research at an event hosted Tuesday by the Radcliffe Institute for Advanced Study.