NVIDIA Builds AI That Creates 3D Objects for Virtual Worlds

The massive virtual worlds created by growing numbers of companies and creators could be more easily populated with a diverse array of 3D buildings, vehicles, characters and more — thanks to a new AI model from NVIDIA Research.

Trained using only 2D images, NVIDIA GET3D generates 3D shapes with high-fidelity textures and complex geometric details. These 3D objects are created in the same format used by popular graphics software applications, allowing users to immediately import their shapes into 3D renderers and game engines for further editing.

The generated objects could be used in 3D representations of buildings, outdoor spaces or entire cities, designed for industries including gaming, robotics, architecture and social media.

GET3D can generate a virtually unlimited number of 3D shapes based on the data it’s trained on. Like an artist who turns a lump of clay into a detailed sculpture, the model transforms numbers into complex 3D shapes.

With a training dataset of 2D car images, for example, it creates a collection of sedans, trucks, race cars and vans. When trained on animal images, it comes up with creatures such as foxes, rhinos, horses and bears. Given chairs, the model generates assorted swivel chairs, dining chairs and cozy recliners.

“GET3D brings us a step closer to democratizing AI-powered 3D content creation,” said Sanja Fidler, vice president of AI research at NVIDIA, who leads the Toronto-based AI lab that created the tool. “Its ability to instantly generate textured 3D shapes could be a game-changer for developers, helping them rapidly populate virtual worlds with varied and interesting objects.”

[…]

GET3D can instead churn out some 20 shapes a second when running inference on a single NVIDIA GPU — working like a generative adversarial network for 2D images, while generating 3D objects. The larger, more diverse the training dataset it’s learned from, the more varied and detailed the output.

NVIDIA researchers trained GET3D on synthetic data consisting of 2D images of 3D shapes captured from different camera angles. It took the team just two days to train the model on around 1 million images using NVIDIA A100 Tensor Core GPUs.

[…]

GET3D gets its name from its ability to Generate Explicit Textured 3D meshes — meaning that the shapes it creates are in the form of a triangle mesh, like a papier-mâché model, covered with a textured material. This lets users easily import the objects into game engines, 3D modelers and film renderers — and edit them.

Once creators export GET3D-generated shapes to a graphics application, they can apply realistic lighting effects as the object moves or rotates in a scene. By incorporating another AI tool from NVIDIA Research, StyleGAN-NADA, developers can use text prompts to add a specific style to an image, such as modifying a rendered car to become a burned car or a taxi, or turning a regular house into a haunted one.

[…]

Source: NVIDIA AI Research Helps Populate Virtual Worlds With 3D Objects | NVIDIA Blog

DNA nets capture COVID-19 virus in low-cost rapid-testing platform


Tiny nets woven from DNA strands cover the spike proteins of the virus that causes COVID-19 and give off a glowing signal in this artist’s rendering. Credit: Xing Wang, University of Illinois

Tiny nets woven from DNA strands can ensnare the spike protein of the virus that causes COVID-19, lighting up the virus for a fast-yet-sensitive diagnostic test—and also impeding the virus from infecting cells, opening a new possible route to antiviral treatment, according to a new study.

Researchers at the University of Illinois Urbana-Champaign and collaborators demonstrated the DNA nets’ ability to detect and impede COVID-19 in human cell cultures in a paper published in the Journal of the American Chemical Society.

“This platform combines the sensitivity of PCR and the speed and low cost of antigen tests,” said study leader Xing Wang, a professor of bioengineering and of chemistry at Illinois. “We need tests like this for a couple of reasons. One is to prepare for the next pandemic. The other reason is to track ongoing viral epidemics—not only coronaviruses, but also other deadly and economically impactful viruses like HIV or influenza.”

DNA is best known for its genetic properties, but it also can be folded into custom nanoscale structures that can perform functions or specifically bind to other structures much like proteins do. The DNA nets the Illinois group developed were designed to bind to the coronavirus spike protein—the structure that sticks out from the surface of the virus and binds to receptors on to infect them. Once bound, the nets give off a fluorescent signal that can be read by an inexpensive handheld device in about 10 minutes.

The researchers demonstrated that their DNA nets effectively targeted the spike protein and were able to detect the virus at very low levels, equivalent to the sensitivity of gold-standard PCR tests that can take a day or more to return results from a clinical lab.

The technique holds several advantages, Wang said. It does not need any special preparation or equipment, and can be performed at , so all a user would do is mix the sample with the solution and read it. The researchers estimated in their study that the method would cost $1.26 per test.

“Another advantage of this measure is that we can detect the entire virus, which is still infectious, and distinguish it from fragments that may not be infectious anymore,” Wang said. This not only gives patients and physicians better understanding of whether they are infectious, but it could greatly improve community-level modeling and tracking of active outbreaks, such as through wastewater.

In addition, the DNA nets inhibited the virus’s spread in live cell cultures, with the antiviral activity increasing with the size of the DNA net scaffold. This points to DNA structures’ potential as therapeutic agents, Wang said.

“I had this idea at the very beginning of the pandemic to build a platform for testing, but also for inhibition at the same time,” Wang said. “Lots of other groups working on inhibitors are trying to wrap up the entire virus, or the parts of the virus that provide access to antibodies. This is not good, because you want the body to form antibodies. With the hollow DNA net structures, antibodies can still access the virus.”

The DNA net platform can be adapted to other viruses, Wang said, and even multiplexed so that a single test could detect multiple viruses.

“We’re trying to develop a unified technology that can be used as a plug-and-play platform. We want to take advantage of DNA sensors’ high binding affinity, low limit of detection, low cost and rapid preparation,” Wang said.

The paper is titled “Net-shaped DNA nanostructures designed for rapid/sensitive detection and potential inhibition of the SARS-CoV-2 .”


More information: Neha Chauhan et al, Net-Shaped DNA Nanostructures Designed for Rapid/Sensitive Detection and Potential Inhibition of the SARS-CoV-2 Virus, Journal of the American Chemical Society (2022). DOI: 10.1021/jacs.2c04835

Source: DNA nets capture COVID-19 virus in low-cost rapid-testing platform

Fitbit accounts are being replaced by Google accounts

New Fitbit users will be required to sign-up with a Google account, from next year, while it also appears one will be needed to access some of the new features in years to come.

Google has been slowly integrating Fitbit into the fold since buying the company back in November 2019. Indeed, the latest products are now known as “Fitbit by Google”. However, as it currently stands, device owners have been able to maintain separate accounts for Google and Fitbit accounts.

Google has now revealed it is bringing Google Accounts to Fitbit in 2023, enabling a single login for both services. From that point on, all new sign ups will be through Google. Fitbit accounts will only be supported until 2025.

From that point on, a Google account will be the only way to go. To aid the transition, once the introduction of Google accounts begins, it’ll be possible to move existing devices over while maintaining all of the recorded data.

[…]

“We’ll be transparent with our customers about the timeline for ending Fitbit accounts through notices within the Fitbit app, by email, and in help articles.”

Whether that will be enough to assuage the concerns of the Fitbit user base – who didn’t have a say on whether Google bought their personal fitness data – remains to be seen.

Source: Fitbit accounts are being replaced by Google accounts | Trusted Reviews

So wonderful cloud – first of all, why should this data go to the cloud anyway? Second, you thought you were giving it to one provider but it turns out you’re giving it to another with no opt-out other than trashing an expensive piece of hardware.

Tiny swimming robots treat deadly pneumonia in mice

Nanoengineers at the University of California San Diego have developed microscopic robots, called microrobots, that can swim around in the lungs, deliver medication and be used to clear up life-threatening cases of bacterial pneumonia.

In mice, the microrobots safely eliminated pneumonia-causing bacteria in the lungs and resulted in 100% survival. By contrast, untreated mice all died within three days after infection.

The results are published Sept. 22 in Nature Materials.

The microrobots are made of algae cells whose surfaces are speckled with antibiotic-filled nanoparticles. The algae provide movement, which allows the microrobots to swim around and deliver antibiotics directly to more bacteria in the lungs. The nanoparticles containing the antibiotics are made of tiny biodegradable polymer spheres that are coated with the cell membranes of neutrophils, which are a type of white blood cell. What’s special about these cell membranes is that they absorb and neutralize inflammatory molecules produced by bacteria and the body’s immune system. This gives the microrobots the ability to reduce harmful inflammation, which in turn makes them more effective at fighting lung infection.

[…]

The team used the microrobots to treat mice with an acute and potentially fatal form of pneumonia caused by the bacteria Pseudomonas aeruginosa. This form of pneumonia commonly affects patients who receive mechanical ventilation in the intensive care unit. The researchers administered the microrobots to the lungs of the mice through a tube inserted in the windpipe. The infections fully cleared up after one week. All mice treated with the microrobots survived past 30 days, while untreated mice died within three days.

Treatment with the microrobots was also more effective than an IV injection of antibiotics into the bloodstream. The latter required a dose of antibiotics that was 3000 times higher than that used in the microrobots to achieve the same effect. For comparison, a dose of microrobots provided 500 nanograms of antibiotics per mouse, while an IV injection provided 1.644 milligrams of antibiotics per mouse.

The team’s approach is so effective because it puts the medication right where it needs to go rather than diffusing it through the rest of the body.

[…]

the researchers say that this approach is safe. After treatment, the body’s immune cells efficiently digest the algae, along with any remaining nanoparticles. “Nothing toxic is left behind,” said Wang.

[…]

Source: Tiny swimming robots treat deadly pneumonia i | EurekAlert!

Journal: Nanoparticle-modified microrobots for in vivo antibiotic delivery to treat acute bacterial pneumonia | nature materials