Nvidia's AI Advances From 3D Simulation to Climate Science

The global AI competition is occurring against a complex backdrop of international AI regulation efforts.

With the EU’s AI Act setting precedents for governance, the US and China are pursuing different regulatory approaches.

It’s well known that the development of AI applications has become increasingly critical for addressing global challenges, from climate change to healthcare accessibility - however - these developments have raised questions about computational efficiency and environmental impact.

Amongst these challenges and successes, Nvidia has positioned itself even further at the front of this evolution.

The company's research division, comprising several hundred scientists and engineers worldwide, has focused on developing practical applications that demonstrate AI's potential beyond conventional uses in data processing and image generation.

In this context, Nvidia’s breakthroughs in 2024 demonstrate the shift from theoretical AI capabilities to practical implementations across multiple sectors, including healthcare, climate science and creative industries. 

Nvidia: pushing the boundaries of Gen AI

Some of Nvidia’s most exciting developments have been in the field of Gen AI.

One of its achievements that stood out last year was the launch of Jetson Nano Super: an affordable Gen AI supercomputer aimed at making advanced AI development more accessible. 

Jensen Huang, the CEO of Nvidia, summarises what he’s excited about with the new gadget: “This is a brand new Jetson Nano Super, offering almost 70 trillion operations per second, 25 Watts and US$249. It runs everything that the HGX does – it even runs LLMs and I can’t wait for all of you to try it. It’s available everywhere.”

Another development in Gen AI is Nvidia’s research that has extended beyond traditional text and image generation to include 3D models, music and sounds.

A key innovation is ConsiStory, a collaboration between Nvidia and Tel Aviv University.

This technology enables the generation of multiple images featuring a consistent main character, a capability crucial for storytelling applications such as comic strips or storyboards.

Nvidia’s researchers claim that their approach, using a technique called subject-driven shared attention, reduces the time required to generate consistent imagery from 13 minutes to approximately 30 seconds.

Another significant development is Edify 3D, a foundation model that allows developers and content creators to rapidly generate 3D objects for prototyping ideas and populating virtual worlds.

This tool is part of Nvidia's broader multimodal architecture for developing visual Gen AI.

Meanwhile, in the audio domain, Nvidia unveiled Fugatto, a foundational Gen AI model that can create or transform any mix of music, voices and sounds based on text or audio prompts.

This technology has potential applications in music production, advertising, video game development and language learning tools.

Advancing scientific applications of AI

Nvidia's research has also yielded promising results in scientific domains.

In climate science, for example, the company introduced StormCast, a Gen AI model for emulating atmospheric dynamics.

This model achieves a spatial resolution of 3 kilometres and an hourly temporal scale, surpassing other machine learning models trained on global data.

According to Nvidia, when applied with precipitation radars, StormCast offers forecasts with lead times of up to six hours that are up to 10% more accurate than the US National Oceanic and Atmospheric Administration's state-of-the-art 3-kilometre regional weather prediction model.

Furthermore, Nvidia’s inception startup Vibrant Planet, is a company that uses AI to inform a wildfire protection collaboration software platform.

“Our mission at Vibrant Planet is to accelerate community and landscape resilience to the ongoing wildfire crisis.

“We do that by developing the best in class AI platform, to estimate forest structure and use that information to inform a wildfire protection collaboration software platform”, says Dr. Tony Chang, Machine Learning Engineering Manager at Vibrant Planet.

Meanwhile, in healthcare, Nvidia collaborated with researchers from the Weizmann Institute of Science and Tel Aviv-based startup Pheno.AI to develop GluFormer, an AI model that predicts an individual's future glucose levels and other health metrics based on past glucose monitoring data.

The researchers report that this model can also predict how a person's glucose levels will respond to specific foods and dietary changes, enabling precision nutrition.

Implications for industry and society

These advancements in AI technology have far-reaching implications across various sectors.

For instance, the improvements in 3D modelling and simulation could revolutionise industries such as video game development, architectural design and virtual reality experiences.

The progress in climate modelling and healthcare applications also demonstrates AI's potential to address pressing global challenges.

More accurate weather predictions could enhance disaster preparedness, while personalised health forecasting could lead to more effective preventive care strategies.

Nvidia's AI technologies are additionally being applied to solve real-world problems.

For example, Illinois-based startup SimBioSys has created TumorSight Viz, a technology that converts MRI images into 3D models of breast tissue, helping surgeons better treat breast cancers.

Meanwhile, in environmental conservation, Clearbot, an AI-driven sea-cleaning boat developed by University of Hong Kong graduates, is autonomously collecting trash from waters in Hong Kong and India.

Explore the latest edition of Technology Magazine and be part of the conversation at our global conference series, Tech & AI LIVE.

Discover all our upcoming events and secure your tickets today.

Technology Magazine is a BizClik brand