We strive to inform readers not only about events in the bitcoin industry, but also about developments in adjacent technology fields — cyber security, and now in the world of artificial intelligence (AI).
ForkLog AI has compiled the most important AI news of the past week.
- Russian bailiffs have been allowed to use AI for debt collection.
- Diia has launched testing of the updated facial recognition system.
- Russian researchers have developed an autonomous version of the LADA Vesta.
- Microsoft and Qualcomm will develop a chip for augmented-reality glasses.
- Nvidia Canvas has added GauGAN2 to generate photorealistic landscapes.
In Russia, debt collection will be handled by artificial intelligence
Russian bailiffs have been allowed to use AI-based systems for debt collection. The relevant law will come into force on 20 June 2022.
Under the document, the algorithms will identify defaulters of administrative fines, including offending drivers, file cases against them, and trace their accounts in banks. AI will also be able to remove people from the “do-not-leave” list after they settle their debts.
The Association of Lawyers of Russia said that, according to the ministry’s digital development plans, by 2023 80% of enforcement service decisions should be automated.
In Diia, beta testing of the updated facial recognition system has begun
The Ministry of Digital Transformation of Ukraine has launched beta testing of the PhotoID facial recognition technology in the Diia app.
The facial-verification feature no longer requires users to turn their heads in different directions during the creation or use of a digital signature.
Anyone may join the PhotoID beta. To do so, one must possess a biometric or foreign passport and register on the Diia site.
In Russia, NAM I developed an autonomous LADA Vesta
The Russian state institute NAM I developed an autonomous car based on the LADA Vesta. It is already being tested in urban conditions.
The robotic car is equipped with radars, cameras and lidars, which operate in concert with detailed digital maps of the surrounding environment.
According to NAM I, the aim of the project is to create an autonomous system intended for integration into a variety of vehicles.
Elon Musk: Tesla Bot could develop a distinct personality
Tesla Bot will be more than a mere humanoid robot for performing useful tasks. It could potentially develop a unique personality and be a companion, says Elon Musk.
According to the billionaire, the robot could become a “incredible friend like C-3PO or R2-D2”.
“Over time, it could develop a unique personality. It doesn’t look like all robots are the same. This personality could evolve to suit the owner, or as you may call it,” said Musk.
Media: Google develops new AR glasses
Google is actively developing new augmented-reality glasses.
According to The New York Times, the company is leveraging work from the Canadian startup North, acquired in June 2020. At the time of the deal’s closure, the company was preparing to release Focals 2.0 smart glasses with a holographic projector capable of displaying various interfaces in front of the user’s eyes.
There is no information yet on an announced release. The publication suggests Google won’t delay a presentation, as a similar product from Apple is reportedly due in the second half of 2022.
Microsoft and Qualcomm to develop a chip for AR glasses
Chipmaker Qualcomm announced a collaboration with Microsoft to develop a chip for augmented-reality glasses.
“We are announcing the development of a custom Snapdragon chip for energy-efficient, ultra-light AR glasses for the Microsoft ecosystem,” said the head of the company, Cristiano Amon.
According to him, the platform will integrate software from both companies: Microsoft Mesh and Qualcomm Snapdragon Spaces XR.
Amon stressed that the new chip will enable very lightweight glasses suitable for a range of mixed-reality tasks.
Scientists have created a quantum neural network capable of learning
A team of American and British researchers has determined the possibility of using convolutional neural networks for data analysis with quantum computers.
According to the researchers, unlike other architectures, quantum convolutional neural networks are immune to the “barren plateau”. This is especially relevant as data dimensionality increases, the researchers said.
The authors also presented a graph-based methodology that they used to determine the neural network’s sensitivity to the “barren plateau.” The researchers believe it will be useful in other applications.
Nvidia updated the tool that turns sketches into landscapes
Nvidia has replaced the original GauGAN model with GauGAN2 in the Nvidia Canvas landscape-generation tool and added a host of new features.
The developers added several brush variants, such as ‘flower’ and ‘straw’. AI has become more advanced to help fill users’ creative gaps.
Nvidia Canvas is available to all, but using it requires an RTX-series GPU with tensor cores.
OpenAI unveils a new algorithm for creating images from descriptions
Engineers at the nonprofit OpenAI have developed the GLIDE algorithm for creating and editing images from text descriptions. It is available on GitHub.
The diffusion-model-based system distorts the original image and adds Gaussian noise. It then triggers the reverse process, gradually removing defects and generating new objects.
The developers used a neural network with 3.5 billion parameters, testing it with the CLIP classifier and independently. They found that the second approach was more effective.
GLIDE generates details in a matter of seconds, including shadows and reflections. It can work with photographs and paintings, adapting to an artist’s style and the “character” of each canvas.
Most notable AI deals of the past week.
From January 3 to 8, 2022, AI startups raised more than $1.08 billion. Here are the most notable deals.
- Avataar raised $45 million to scale AI technology for life-sized 3D product modeling.
- Pendulum raised $5.9 million to develop a service for tracking harmful narratives on the internet.
- RoadRunner Recycling raised $70 million.
- iProov raised $70 million to enhance facial recognition technology.
- ReadyWorks raised $8 million to expand a service for managing corporate infrastructure.
- Fractal Analytics raised $360 million to enhance tools for analytics of big data.
- Formant raised $18 million to scale the platform for managing robot fleets.
- Creators of the online collaboration workspace Miro raised $400 million.
- Human Security raised $100 million to develop tools for bot detection.
- AIMMO raised $12 million to advance data-labeling technology.
Also on ForkLog:
- In China developed the world’s first AI prosecutor.
- In the US, urged to ban the “killer robots” from autonomously eliminating people.
- Voice assistant advised a child to perform a deadly challenge.
What to read this weekend?
In this piece, the ForkLog AI editors explain how the AI industry changed in 2021 and what to expect in the future.
Subscribe to ForkLog AI news on Telegram: ForkLog AI — all the AI news!