Goto

Collaborating Authors

MIT Technology Review


The Download: AI to predict ice, and healthcare censorship in China

MIT Technology Review

The news: Researchers have used deep learning to model more precisely than ever before how ice crystals form in the atmosphere. Their paper, published this week in PNAS, hints at the potential to significantly increase the accuracy of weather and climate forecasting. How they did it: The researchers used deep learning to predict how atoms and molecules behave. First, models were trained on small-scale simulations of water molecules to help them predict how electrons in atoms interact. The models then replicated those interactions on a larger scale, with more atoms and molecules.


Deep learning can almost perfectly predict how ice forms

MIT Technology Review

"The properties of matter emerge from how electrons behave," says Pablo Piaggi, a research fellow at Princeton University and the lead author on the study. "Simulating explicitly what happens at that level is a way to capture much more rich physical phenomena." It's the first time this method has been used to model something as complex as the formation of ice crystals, also known as ice nucleation. This is one of the first steps in the formation of clouds, which is where all precipitation comes from. Xiaohong Liu, a professor of atmospheric sciences at Texas A&M University who was not involved in the study, says half of all precipitation events--whether snow or rain or sleet--begin as ice crystals, which then grow larger and result in precipitation.


How to craft effective AI policy

MIT Technology Review

Jennifer Strong: The applications of artificial intelligence are so embedded in our everyday lives it's easy to forget it's there… But these systems, like ones powering Instagram filters or the price of a car ride home… can rely on pre-existing datasets that fail to paint a complete picture of consumers. It means people become outliers in that data – often the same people who've historically been marginalized. It's why face recognition technologies are least accurate on women of color, and why ride-share services can actually be more expensive in low-income neighborhoods. So, how do we stop this from happening? Well would you believe a quote from Harry Potter and his wizarding world… might create a good starting point for this conversation? I'm Jennifer Strong and this episode, our producer Anthony Green brings you a conversation about equity from MIT Technology Review's A-I conference, EmTech Digital.


The Download: experimental embryos and the US monkeypox emergency

MIT Technology Review

In a search for novel forms of longevity medicine, a biotech company based in Israel says it intends to create embryo-stage versions of people in order to harvest tissues for use in transplant treatments. The company, Renewal Bio, is pursuing recent advances in stem-cell technology and artificial wombs, demonstrated by Jacob Hanna, a biologist at the Weizmann Institute of Science in Rehovot. Earlier this week, Hanna showed that starting with mouse stem cells, his lab could form highly realistic-looking mouse embryos and keep them growing in a mechanical womb for several days until they developed beating hearts, flowing blood, and cranial folds. It's the first time such an advanced embryo has been mimicked without sperm, eggs, or even a uterus. Now Hanna has set his sights on extending the technology to humans--he's already experimenting with human cells and hopes to eventually produce artificial models of human embryos.


Automated techniques could make it easier to develop AI

MIT Technology Review

Although automated machine learning has been around for almost a decade, researchers are still working to refine it. Last week, a new conference in Baltimore--which organizers described as the first international conference on the subject--showcased efforts to improve autoML's accuracy and streamline its performance. There's been a swell of interest in autoML's potential to simplify machine learning. Companies like Amazon and Google already offer low-code machine-learning tools that take advantage of autoML techniques. If these techniques become more efficient, it could accelerate research and allow more people to use machine learning. The idea is to get to a point where people can choose a question they want to ask, point an autoML tool at it, and receive the result they are looking for.


The Download: a big DeepMind breakthrough, and fixing the US grid

MIT Technology Review

The news: DeepMind says its AlphaFold tool has successfully predicted the structure of nearly all proteins known to science. It's a massive boost to the existing database of 1 million proteins it released last year, and includes structures for plants, bacteria, animals, and many other organisms. Why it matters: The expanded database opens up huge opportunities for AlphaFold to have impact on important issues such as sustainability, fuel, food insecurity, and neglected diseases, according to Demis Hassabis, DeepMind's founder and CEO. Scientists could use the findings to better understand diseases, and to speed innovation in drug discovery and biology, he added. AI for protein folding represents such a major advance that it was chosen as one of MIT Technology Review's 10 Breakthrough Technologies this year.


DeepMind has predicted the structure of almost every protein known to science

MIT Technology Review

However, for many proteins "we're interested in understanding how their structure is altered by mutations and natural allelic variation, and that won't be addressed by this database," said AlQuraishi. "But of course the field is developing fast, and so I expect tools to accurately model protein variants will begin to appear soon," he added. The quality of AlphaFold's predictions may also not be as accurate for rarer proteins with less available evolutionary information, says Peng. The move is the latest development in DeepMind's push into "digital biology," where "AI and computational methods can help to understand and model important biological processes," said Hassabis. Hassabis also leads a new venture, also owned by Alphabet, called Isomorphic Labs, which is developing AI for drug discovery. Pushmeet Kohli, head of AI for science at DeepMind, said the company has plenty of challenges in the life sciences it still wants to tackle, such as how proteins behave and interact with other proteins.


The Download: Chinese robotaxi drivers, and AI gun detection

MIT Technology Review

When Liu Yang started his current job, he found it hard to go back to driving his own car: "I instinctively went for the passenger seat. Or when I was driving, I would expect the car to brake by itself," says the 33-year-old Beijing native, who joined the Chinese tech giant Baidu in January 2021 as a robotaxi driver. Robotaxi driver is an occupation that only exists in our time, the result of an evolving technology that's advanced enough to get rid of a driver--most of the time, in controlled environments-- but not good enough to convince authorities that they can do away with human intervention altogether. Liu is one of the hundreds of safety operators employed by Baidu, "driving" five days a week in Shougang Park. But despite having only worked for the company for 19 months, he already has to think about his next career move, as his job will likely be eliminated within a few years.


A day in the life of a Chinese robotaxi driver

MIT Technology Review

Robotaxi safety operator is an occupation that only exists in our time, the result of an evolving technology that's advanced enough to get rid of a driver--most of the time, and in controlled environments-- but not good enough to convince authorities that they can do away with human intervention altogether. Today, self-driving companies from the US, Europe, and China are racing to bring the technology to commercial application. Most of them, including Apollo, the self-driving arm of Baidu, have started on-demand robotaxi trials on public roads yet still need to operate with various constraints. With an associate degree in human resources, Liu has no academic training related to this job, But he has always loved driving, and he acted as the driver for his boss in a previous role. When he heard about the self-driving technologies, his curiosity pushed him to look up related jobs online and apply.


A digital human could be your next favorite celebrity--or financial advisor

MIT Technology Review

"Rising demand is driving the boom of digital humans," says Shiyan Li, head of the digital human and robotics business at Baidu, which created the digital model-actor, Gong. "In China alone, there are over 400 million ACGN (animation, comics, games, and novel) fans, and an enterprise market worth hundreds of billions of dollars centered on digital humans." And according to a company that tracks business registrations, Qichacha, China now has more than 280,000 enterprises that engage in digital human-related activities. The debut of Baidu's digital celebrity may not seem like much at first, as the concept of "virtual idols" has been around for years. For example, US virtual influencer Lil Miquela has been appearing alongside real human celebrities in online advertisements and TV commercials since 2016, gaining over three million Instagram followers.