Google Deep Mind has developed an artificial intelligence algorithm that can play games like chess and Go. Yeah, ok, this isn’t new info. We’ve been battling AI in complex abstract strategy games for years, beginning famously with Deep Blue vs. Gary Kasperov back in the mid-90s.

However, the gaming part isn’t the talking point here – that’s not the only thing Deep Mind does. As recently discovered during development it can also perform sorting which, when translated to traditional C++ code, outperforms previous sorting algorithms developed by humans throughout the development of computer science. You read that right – Deep Mind developed its own sorting algorithm, and it’s faster than anything we’ve created ourselves.

DeepMind AI’s self-developed sorting algorithm could potentially offer new insights into sorting algorithms which can outperform traditional computer algorithms like Quick sort.

Sorting algorithms have been a focus for computer scientists for decades. From bubble sort and selection sort to more advanced programs like merge and quick sort. An improved sorting algorithm means not only faster data sorting, but also more efficient use of resources, ultimately allowing greater capacity for other tasks. But, with rapid development of Google DeepMind AI based in London, traditional sorting algorithms may soon be retired.

The new algorithm, known as AlphaDev, allowed DeepMind to test new ways of sorting elements. The team built the algorithm on top of the technology used for its AI algorithm AlphaZero which has shown success in mastering games of chess, shogi and Go.

The new experiment, featured in a research paper published in Nature, is already part of two C++ libraries and is already being used by developers all over the world.

Sorting Numbers

Initially, DeepMind AI sorting algorithm was used to sort simple numbers by their size.

Initially, the program was only tasked with sorting 3 to 5 numbers at a time. This is important because AlphaDev operates at an assembly level of instructions.

That means that the code is generated by automated compilers of code humans write such as C++ and is then translated into machine language – 1s and 0s.

AlphaDev is more successful at sorting quick sequences of number as it achieves 70% increased efficiency.

It is also 1.7% more efficient for sequences that surpass 250,000 elements as opposed to sorting algorithms that are part of the C++ standard libraries.

In other words, when a user submits a sorting query, DeepMind AI sorting algorithm will conduct a faster sorting process, which can save both time and resources once it’s employed on large scale projects.

Reinforcement Learning

AlphaDev can compare values, move values between different locations or jump to a different instruction of a program.

Every time the algorithm takes a new step, it will try to sort a set of list and for each successful sort, it’ll be rewarded, thanks to the machine learning algorithm known as reinforcement learning.

The system will play this game with rewards until all elements are sorted correctly, or the program length limit has been reached. In case of the latter, the game is restarted and the algorithm starts sorting again.

DeepMind AI sorting algorithm also tried to solve algorithms that are not related to sorting.

It was also used on hash data structures, and tested to convert data from one format into bytes which took 67% less time compared to a human-written version of the algorithm.

The researchers believe that DeepMind AI sorting algorithm could be applied to more problems, even into the hardware design code, Daniel Mankowitz, a computer scientist at DeepMind who led the research said in a statement to Nature. “We really want to be tackling the whole stack.”

Nvidia’s research department recently unveiled an eye-opening new AI model that can transform dull 2D video clips into detailed 3D graphics called Neuralangelo. Nvidia Neuralangelo is based on neural networks and can successfully generate lifelike constructions, buildings, and other realistic imagery based solely on 2D videos. With this research, the company hopes to recreate virtual replicas of different structures including historical monuments, buildings, and sculptures.

Nvidia is forging a solid path through the AI industry in 2023, cementing itself as a global leader in the industry. Recently, it was announced that Nvidia’s AI-driven chipsets generates billions of dollars in gevenue for the company, and now a new AI model powered by neural networks promises advancements in architecture and 3D modeling on a level we haven’t seen since the early days of digital design.

Neuralangelo is quite obviously inspired by the famous Italian artist and sculptor Michelangelo, whose crafty hands were able to sculpt complex works of art from materials such as marble. Similarly, Nvidia Neuralangelo can generate detailed and intricate shapes and textures just based on simple 2D videos.

Nvidia Neuralangelo: 2D to 3D

Developing 3D models from scratch using traditional methods is extremely time-consuming and challenging. However, with this new AI-powered model, designers can generate structures such as “roof shingles, panes of glass and smooth marble” with impeccable details from 2D videos not only faster than traditional non-AI approaches but also more accurately than other AI-powered approaches.

All you need is a smartphone video, and Nvidia Neuralangelo will turn it into accurate renditions of your preference without sacrificing your time, money, and resources. This approach can help with landscape and architectural design, and video gaming for both corporate and personal needs.

“The 3D reconstruction capabilities Neuralangelo offers will be a huge benefit to creators, helping them recreate the real world in the digital world,” Ming-Yu Liu, senior director of research and co-author on the paper said in a statement.

He also added that developers will be able to import detailed objects into virtual environments for 3D modeling and video games starting from small objects like sculptures to massive skyscrapers.

Neural Rendering

Previously used AI models that could transform and reconstruct 3D objects had several issues including redundancy and struggling with homogenous and strong color textures. However, Nvidia Neuralangelo uses instant neural graphics primitives which helps generate intricate details and textures.

A 2D video of a scene or an object that needs to be transformed is fed into the system that selects the frames that can capture several viewpoints. This approach is similar to how different artists focus on objects they are trying to sculpt or paint from different angles.  

Once the system determines the camera angle of each frame that needs transforming, Nvidia Neuralangelo transforms it into a rough 3D presentation, after which the model sharpens the details and optimises the entire 3D structure.

Neuralangelo will be presented at the Conference on Computer Vision and Pattern Recognition (CVPR) along with other Nvidia Research projects on June 18-22.

OpenAI and its ChatGPT generative AI model kickstarted a whole AI revolution in 2023. Every day, we see that new AI tools are making it to the light of day and showcasing great features that can make our everyday life easier and more productive. Still, the next step OpenAI allegedly plans to take may lead to even greater revolutionization of AI in the world. According to a new report, OpenAI is working on an open-source AI model that could change everything.

According to a new report published in The Information, OpenAI, the company behind the excellent ChatGPT generative AI model, is working on releasing a new language model which would be open-source.

Open-source programs have existed for years, and essentially it meant that the public would have access to the source code behind those programs. With OpenAI releasing an open-source AI model, people would get a chance to see the code behind ChatGPT and other GPT models.

Not the Only Open-Source AI Model

The reasoning behind this step, according to the report, is to fight competition. OpenAI’s AI models haven’t been available to the public since 2019, when the GPT-2 model was released, and caused a lot of concern regarding the possibility of it being used to spread propaganda and generate fake news that could lead to misinformation and panic.

One such concern was already voiced by The Guardian’s report in 2019, highlighting that the spread of fake information and propaganda could cause a new political threat.

Nevertheless, according to this report, it is believed that OpenAI is open-sourcing its new AI model due to competition with Meta, which also released an open-source AI model called LLaMa earlier this year.

The Meta’s AI model is quite sophisticated and advanced, being trained with over 65 billion parameters. Its complexity could put enough pressure on both Google and OpenAI to either open-source their models or release alternative models with their code visible to the public.

In return, to keep up with the competition, OpenAI would need to make better software and offer more features for its users. That said, the potentially new AI model won’t be a direct rival to GPT-4 and ChatGPT in particular.

Right now, ChatGPT based on GPT-3 model is available for free, while the more premium version of the program costs $20/month and runs on GPT-4 language model. That said, if OpenAI published its open-source AI model, it’ll likely offer fewer features than GPT-4 so it wouldn’t reduce the commercial value of ChatGPT Premium.

The potential release of the open-source AI model could change the way different tech companies develop their AI products, especially Google. After all, last month Google merged its two main AI divisions into one, so we can expect innovative AI products coming from all sides.