Lifeboat Foundation

Researchers from Tsinghua University, China, have developed an all-analog photoelectronic chip that combines optical and electronic computing to achieve ultrafast and highly energy-efficient computer vision processing, surpassing digital processors.

Computer vision is an ever-evolving field of artificial intelligence focused on enabling machines to interpret and understand visual information from the world, similar to how humans perceive and process images and videos.

It involves tasks such as image recognition, object detection, and scene understanding. This is done by converting analog signals from the environment into digital signals for processing by neural networks, enabling machines to make sense of visual information. However, this analog-to-digital conversion consumes significant time and energy, limiting the speed and efficiency of practical neural network implementations.

For the first time, a physical neural network has successfully been shown to learn and remember “on the fly,” in a way inspired by and similar to how the brain’s neurons work.

The result opens a pathway for developing efficient and low-energy machine intelligence for more complex, real-world learning and memory tasks.

Published today in Nature Communications, the research is a collaboration between scientists at the University of Sydney and University of California at Los Angeles.

A little more than two years have passed since the Lucy mission launched on an Atlas V rocket, ultimately bound for asteroids that share an orbit with Jupiter. After a gravity assist from Earth in 2022, the spacecraft has been making a beeline for an intermediate target, and now it is nearly there.

On Wednesday, the $1 billion mission is due to make its first asteroid flyby, coming to within 265 miles (425 km) of the small main belt asteroid Dinkinesh. In a blog post, NASA says the encounter will take place at 12:54 pm ET (16:54 UTC).

About an hour before the encounter, the spacecraft will begin attempting to lock on to the small asteroid so that its instruments are oriented toward it. This will allow for the best possible position to take data from Dinkinesh as Lucy speeds by at 10,000 mph (4,470 meters per second).

Tesla is ramping up hiring for its humanoid robot program, Optimus, including some reinforcement learning engineers.

It was hard to take Tesla Bot seriously when Elon Musk announced it by having someone dressed as a robot dance on stage.

To this day, many people are taking the project seriously, but Tesla certainly is.

The search is on worldwide to find ways to extract carbon dioxide from the air or from power plant exhaust and then make it into something useful. One of the more promising ideas is to make it into a stable fuel that can replace fossil fuels in some applications. But most such conversion processes have had problems with low carbon efficiency, or they produce fuels that can be hard to handle, toxic, or flammable.

Now, researchers at MIT and Harvard University have developed an efficient process that can convert carbon dioxide into formate, a liquid or solid material that can be used like hydrogen or methanol to power a fuel cell and generate electricity. Potassium or sodium formate, already produced at industrial scales and commonly used as a de-icer for roads and sidewalks, is nontoxic, nonflammable, easy to store and transport, and can remain stable in ordinary steel tanks to be used months, or even years, after its production.

An FDA advisory panel tended to embrace a new gene therapy treatment from Vertex and CRISPR for sickle cell anemia on Tuesday.

President Joe Biden’s executive order plans for standardized digital watermarking rules.

President Joe Biden’s executive order on artificial intelligence is a first-of-its-kind action from the government to tackle some of the technology’s greatest challenges — like how to identify if an image is real or fake.

Among a myriad of other demands, the order, signed Monday, calls for a new set of government-led standards on watermarking AI-generated content. Like watermarks on photographs or paper money, digital watermarks help users distinguish between a real object and a fake one and determine who owns it.

Continue reading “Watermarks aren’t the silver bullet for AI misinformation” »

You can now register.ing domains as part of an early access program. Some aren’t cheap: th.ing, for example, will cost you $38,999.99 per year from GoDaddy.

You can now register.ing domains as part of Google’s early access period, though you’ll have to pay “an additional one-time fee,” Google’s Christina Yeh wrote in a blog post. That fee will go down daily through December 5th. At 16:00 UTC (noon ET) on that date,.ing domains will officially be publicly available.

A few of us at The Verge have had some fun plugging different words into GoDaddy’s domain search to see what.ing domains are up for grabs. Some common words with-ing endings are really expensive; think.ing and buy.ing cost $38,999.99 and $129,999.99, respectively, per year to register. But I’ve found a few good ones that are pretty affordable — good luck on the hunt!

Nearly five years ago, DeepMind, one of Google’s more prolific AI-centered research labs, debuted AlphaFold, an AI system that can accurately predict the structures of many proteins inside the human body. Since then, DeepMind has improved on the system, releasing an updated and more capable version of AlphaFold — AlphaFold 2 — in 2020.

And the lab’s work continues.

Today, DeepMind revealed that the newest release of AlphaFold, the successor to AlphaFold 2, can generate predictions for nearly all molecules in the Protein Data Bank, the world’s largest open access database of biological molecules.