Today in AI: Nvidia's new AI model, EU and US promise voluntary AI act and more

Nvidia's Neuralangelo AI model can create 3D objects from 2D videos

Neuralangelo works by analysing a 2D video clip of an object that has been filmed from various angles. It selects several frames of reference from the video to get an idea of what the object would look like, which allows it to get a sense of depth, size, and shape of the object.

• Once analysed, the AI model then creates a rough 3D representation of the object, and then optimises it by adding details and textures. The resulting 3D model can then be used in, "virtual reality applications, digital twins or robotics development".


• The company says that the AI model can generate "3D structures with intricate details and structures", and these models can then be imported into editing and design applications for further tweaking.

European Union and US to release voluntary code of conduct for AI

The goal is to establish common standards for AI in democracies as the debate on effective regulation of the technology rages on.

• The European Union (EU) and the US have said they will release a voluntary code of conduct for the use of artificial intelligence (AI) and will ask "like-minded countries" to join them.


• European Commission Vice-President, Margrethe Vestager, said that a draft of the act would be put forward "within weeks".


• Vestager hoped to bring, "as many onboard as possible", and in the, "broadest possible circle -- with our friends in Canada, in the UK, in Japan, in India".

MIT researchers develop small-scale language model more efficient than larger ones

The self-learning language model can learn from its own predictions eliminating the use of annotated training data.

• Researchers at MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) have developed a small, self-learning language model that surpasses large scale language models. CSAIL's algorithm for the model called Simple Pseudo-Label Editing (SimPLE) allows it to learn from its own predictions, eliminating the use of annotated training data.


• “Our small model is trained to grasp the core principle of language understanding — contextual entailment, while LLMs do not explicitly learn about it," said Hongyin Luo, postdoctoral associate at MIT CSAIL and lead author on the research.