Published on October 18, 2023, 11:11 am
Generative AI has been taking the robotics world by storm, and it’s no surprise. With various approaches to harnessing this emerging technology, from natural language commands to design, the possibilities are endless. During a recent visit to Nvidia’s South Bay headquarters, Deepu Talla, Vice President and General Manager of Embedded & Edge Computing, shared insight on generative AI.
Talla emphasized the impact of generative AI on productivity. “You can already see the productivity improvement,” he expressed. While it may not be perfect yet, generative AI has the ability to compose an email that jumpstarts the process, saving time and effort. “It’s giving me 70%. There are obvious things you can already see that are definitely a step function better than how things were before.”
Nvidia was preparing to announce its latest developments in this field. The company revealed exciting news during ROSCon, including the availability of Nvidia Isaac ROS 2.0 and Nvidia Isaac Sim 2023 platforms. By incorporating generative AI into these systems, Nvidia aims to accelerate its adoption among roboticists. Notably, over 1.2 million developers have interacted with the Nvidia AI and Jetson platforms, including esteemed clients like AWS, Cisco, and John Deere.
Among these announcements is the introduction of the Generative AI Playground for Jetson—a platform granting developers access to open-source large language models (LLMs). Developers can leverage optimized tools and tutorials for deploying LLMs, diffusion models for interactive image generation, vision language models (VLMs), and vision transformers (ViTs) combining vision AI with natural language processing.
These advancements are particularly significant as they enable systems to make informed decisions even in unfamiliar situations. While simulated environments have their limits when it comes to training autonomous systems, models like those offered by Nvidia provide real-time adjustment capabilities and a more intuitive interface through natural language integration.
“Generative AI will significantly accelerate deployments of AI at the edge with better generalization, ease of use, and higher accuracy than previously possible,” Talla stated. The expansion of Metropolis and Isaac frameworks on Jetson, combined with the power of transformer models and generative AI, addresses this need.
In addition to these developments, the latest versions of Nvidia’s platforms bring improvements in perception and simulation capabilities. With ongoing advancements in generative AI, roboticists can enhance their systems’ abilities to interpret complex scenarios accurately, adapt on the fly, and provide more comprehensive analyses.
Generative AI holds immense potential for robotics, as it equips machines with the tools to understand and interact with their surroundings effectively. With Nvidia at the forefront of this transformative technology, we can expect further breakthroughs that will shape the future of robotics in various industries.