Skip to main content
NVIDIA AI Podcast

Bringing Robots to Life with AI: The Three Computer Revolution - Ep. 274

52 min episode · 2 min read
·
Bringing Robots,Yashraj Narang

Episode

52 min

Read time

2 min

Topics

Fundraising & VC, Design & UX, Artificial Intelligence

AI-Generated Summary

Key Takeaways

  • Three Computer Framework: Modern robotics requires DGX systems for model training, Omniverse plus Cosmos for simulation and world modeling to generate synthetic training data, and Jetson AGX Thor chips for real-time onboard inference, creating an integrated development pipeline.
  • Imitation vs Reinforcement Learning: Imitation learning uses human demonstrations to teach robots human-like behaviors efficiently, while reinforcement learning discovers novel solutions through trial and error, potentially achieving superhuman performance in speed and precision for tasks like assembly.
  • Simulation Data Strategy: Robotics lacks internet-scale training data, making simulation critical. Close the sim-to-real gap through domain randomization of physics parameters and visual properties, domain adaptation to specific environments, or domain invariance by removing unnecessary information from training data.
  • Humanoid Robot Rationale: Humanoid form factors enable robots to operate in human-designed environments without modification, accessing stairs built for human leg dimensions, doors at human heights, and tools like hammers and screwdrivers designed for human hands, accelerating deployment.

What It Covers

Yashraj Narang, head of NVIDIA's Seattle Robotics Lab, explains the three computer revolution enabling intelligent robots: DGX systems for training AI models, Omniverse and Cosmos for simulation and synthetic data generation, and Jetson AGX for onboard inference.

Key Questions Answered

  • Three Computer Framework: Modern robotics requires DGX systems for model training, Omniverse plus Cosmos for simulation and world modeling to generate synthetic training data, and Jetson AGX Thor chips for real-time onboard inference, creating an integrated development pipeline.
  • Imitation vs Reinforcement Learning: Imitation learning uses human demonstrations to teach robots human-like behaviors efficiently, while reinforcement learning discovers novel solutions through trial and error, potentially achieving superhuman performance in speed and precision for tasks like assembly.
  • Simulation Data Strategy: Robotics lacks internet-scale training data, making simulation critical. Close the sim-to-real gap through domain randomization of physics parameters and visual properties, domain adaptation to specific environments, or domain invariance by removing unnecessary information from training data.
  • Humanoid Robot Rationale: Humanoid form factors enable robots to operate in human-designed environments without modification, accessing stairs built for human leg dimensions, doors at human heights, and tools like hammers and screwdrivers designed for human hands, accelerating deployment.

Notable Moment

Narang reveals that neural robot dynamics models can be continuously fine-tuned with real-world data to account for wear and tear, creating self-updating simulators that maintain accuracy as physical robots change over time, enabling perpetual model improvement.

Know someone who'd find this useful?

Episode Transcript

Hello, and welcome to the NVIDIA AI podcast. I'm your host, Noah Kravitz. Our guest today is Yashraj Narang. Yash is a senior research manager at NVIDIA and the head of the Seattle Robotics Lab, which I'm really excited to, learn more about along with you today. Yash's work focuses on the intersection of robotics, AI, and simulation, and his team conducts fundamental and applied research across the full robotics stack, including perception, planning, control, reinforcement learning, imitation learning, simulation, and vision language action models. Full robotics stack, like it says. Prior to joining NVIDIA, Yash completed a PhD in material science and mechanical engineering from Harvard University and a master's in mechanical engineering from MIT. And he's here now to talk about robots, the the field of robotics, robotics learning, all kinds of awesome stuff. I'm so excited to have you here, Yash. So thank you for joining the podcast. Welcome. Thank you so much, Noah. So maybe first things first, and this is a very selfish question I I mentioned before we started, but I I think the listeners will will be into it too. I've never been to the Seattle Robotics Lab. I don't know much about it. Can we start with having you talk a little bit about your own role, your background if you like, and, you know, give us a little peek into what the Seattle Lab is all about? Yeah. Absolutely. So the Seattle Robotics Lab, it started in, I believe, October 2017, and I actually joined the lab in December 2018. And the lab was started by Dieter Fox, who's a professor at University of Washington. And, you know, at the time, I believe he had a conversation with Jensen at a conference. Jensen Huang, of course, is CEO of NVIDIA. And Jensen thinks way far out into the future. And at that point, he was getting really excited about robotics. He said, you know, essentially, that we need a research effort in robotics at NVIDIA, and that's really how the lab started. So that was kind of the birth of the lab. And at the beginning, the lab you know? And it still does a very academic focus. Okay. So we consistently have really high engagement at conferences. We publish a lot. We do a lot of fundamental and applied research. And recently, NVIDIA has been developing, especially over the past few years, a really robust product and engineering effort as well. And so we're working more closely and closely with them to try to get some of our research out into the hands of the community. So, you know, fundamental academic mission, but it's really important for us as well to, transfer our research, and get it out there for everyone to use. Fantastic. And you mentioned Dieter Fox, I believe, at UW, University of Washington. Is the lab is there a relationship there? Yeah. So, when Dieter started the lab, we, you know, over a number of years, had …

Get the full transcript (9,644 words) + summary by email — free

One-time email with the complete transcript and AI summary of this episode. No account needed.

One email, no spam. We’ll also show you what SignalCast does.

Browse all NVIDIA AI Podcast transcripts →

You just read a 3-minute summary of a 49-minute episode.

Get NVIDIA AI Podcast summarized like this every Monday — plus up to 2 more podcasts, free.

Pick Your Podcasts — Free

Keep Reading

Books, tools, and gear mentioned in this episode

SignalCast may earn commission on purchases via these links. As an Amazon Associate, SignalCast earns from qualifying purchases.

Tools

  • by NVIDIA

    Omniverse plus Cosmos for simulation and world modeling to generate synthetic training data
  • by NVIDIA

    DGX systems for model training, Omniverse plus Cosmos for simulation and world modeling to generate synthetic training data

Gear

  • by NVIDIA

    Jetson AGX Thor chips for real-time onboard inference, creating an integrated development pipeline
  • by NVIDIA

    Jetson AGX for onboard inference, creating an integrated development pipeline
  • by NVIDIA

    Modern robotics requires DGX systems for model training, Omniverse plus Cosmos for simulation and world modeling to generate synthetic training data, and Jetson AGX for onboard inference

More from NVIDIA AI Podcast

We summarize every new episode. Want them in your inbox?

Similar Episodes

Related episodes from other podcasts

Explore Related Topics

This podcast is featured in Best AI Podcasts (2026) — ranked and reviewed with AI summaries.

Read this week's AI & Machine Learning Podcast Insights — cross-podcast analysis updated weekly.

You're clearly into NVIDIA AI Podcast.

Every Monday, we deliver AI summaries of the latest episodes from NVIDIA AI Podcast and 192+ other podcasts. Free for one show.

Start My Monday Digest

No credit card · Unsubscribe anytime