7 Interesting Robotics and AI Initiatives From NVIDIA For Engineers!
NVIDIA wants robots everywhere. And wants them to be better and smarter. Big part of the solution is to manufacture more powerful chipsets. This is what the company was historically known for.
But another part is to allow other companies & independent 3rd party engineers to build and train robots easier and faster. How to make this vision come true? Here are some interesting initiatives from NVIDIA that may not have direct impact on consumers in the short term, but paves the way for the company to be a key part of a smart AI-powered future.
1. NVIDIA AI SDK — JetPack
No, it’s actually not a physical jetpack that we’re going to write about here. But just like a physical device NVIDIA Jetpack AI can lift your Artificial Intelligence (AI) projects and propel them through the air!
JetPack is an SDK (Software Development Kit) dedicated to all AI developers and engineers. For those of you who don’t come from a technical world: An SDK or ‘devkit’ is simply a set of software development tools that allows developers to create new applications or enrich them with advanced functionalities — for example Google Maps SDK may be helping mobile developers to create navigation apps or smartphone tourist guide apps. Interesting trivia: in fact when you’re using any mobile app on your phone on average it relies on over 15 different external APIs.
The newest JetPack SDK version is 3.1 and it’s probably the most comprehensive solution for building AI applications available today. And there is one more ultimately cool thing about it — it’s free to use!
For those of you who are more familiar with the world of programming: this SDK includes TensorRT is a high performance deep learning inference runtime for image classification, segmentation, and object detection neural networks. It speeds up deep learning inference as well as reducing the runtime memory footprint for convolutional and deconv neural networks (so could be a great choice for image processing applications).
There’s also Multimedia API package that provides low level APIs for more flexible application development. To put it in a simple way if you’re building a robot that’s relying or camera or other sensors (both for input and output) it may be very helpful.
More about JetPack: https://developer.nvidia.com/embedded/jetpack
It may be a great idea to combine JetPack SDK and Jetson computer in your robotic project.
Jetson is simply AI supercomputer on a module. It’s not exactly kind of computer that sits on your desk. It’s a smaller one and can be used inside devices like robots, drones, smart cameras, and portable medical devices. It can efficiently process complex data on-board those devices enabling them to do more.
So why it might be a good idea to combine both? First of all Jetson is known for its efficiency and incredible computing power. The newest Jetson TX2 has roughly twice the performance of its predecessor the TX1. What does it mean in practice and what are new opportunities? Just think of vision processing and robot interacting with its environment in real-time. Whatever comes to your mind from traffic cameras, manufacturing robotics and smart sensors — it could be taken to the next level
Plus, it’s consuming little energy. It’s a little smart supercomputer that can go a really long way!
More about Jetsons: http://www.nvidia.com/object/embedded-systems-dev-kits-modules.html
2. Isaac — photorealistic, simulated environment for robots and drones
Thanks to this initiative robots will not need the physical world to learn! In future they will learn doing things in their own virtual world. It means it’s going to be fast and inexpensive.
Named after famous sci-fi writer Isaac Asimov. One of the personal favorites by the way.
Isaac is an AI-based software platform that lets developers train their robots using detailed and highly realistic test scenarios. It’s kind of logical continuation: NVIDIA is not only giving you tools to build your robots but also a way to train & test them before real-world deployment.
The initiative was unveiled during GTC conference earlier this year. See NVIDIA CEO Jensen Huang announcing Isaac.
During his keynote, Huang discussed example usage areas of Isaac lab. Now new mechanical systems like robots or drones working within virtualized environment can teach themselves new things. It’s possible to run several instances of it at the same time and see them improved after each training. Once a simulation is complete, the trained neural network of the system can be quickly migrated to real-world robots. Developers can then iterate and tune the robot learning and testing methodology while moving intelligence back and forth between the two environments — virtual and real one.
The lab is photorealistic and obeys the laws of physics. Toyota is amongst the first clients for the Isaac Initiative. It is using the platform for service robots and drone maker brand Teal.
“That’s all I’ve got…” — Huang said towards the end of his Isaac presentation. Funny way to close the presentation considering what they have accomplished!
3. NVIDIA® Jetson™ Developer Challenge — open for all talented Engineers!
NVIDIA wants to recognize the best AI and robotic talents worldwide! New and existing projects are welcomed! So whatever you’ve got you can show it the world now. The only requirement is that you need to leverage Jetson computer in your creation and show innovative use of AI.
Exciting prizes, recognition and opportunity to accelerate your project! All finalists will receive an expense paid trip to GTC in Silicon Valley, California for the opportunity to present their NVIDIA Jetson-enabled creation.
We are super proud and excited that it is hosted on our platform! Check out more details, deadlines and list of eligible countries from here: nvidia.challengerocket.com
Open Challenge for individual engineers, teams and startups.
4. NVIDIA Inception Program
This program is great opportunity to all startups who are revolutionizing industries with advances in AI and data science. Joining it you can get a custom set of ongoing benefits, from hardware grants and marketing support to training with deep learning experts.
Apply here: https://www.nvidia.com/en-us/deep-learning-ai/startups/
5. NVIDIA DriveWorks platform for self-driving car
It’s not a new project but pretty cool one. All Tesla cars shipping today feature the NVIDIA Drive PX 2. Here’s a visualized look at how the AI system sees the world, as well as footage of the NVIDIA BB8 AI car navigating the world.
6. NVIDIA Jetson Reference Platforms
Another initiative to ignite a robotics revolution. Can be used by accelerate robotics research and development.
In the simple words: if you’re a company or individual engineer willing to build a new smart machine there is no need to re-invent the wheel and work from scratch. Use reference model instead. That is enabling developers to create prototypes faster. But it’s not only a time but also about safety. Relying on well-tested and documented reference models can help in that aspect a lot.
More details: https://developer.nvidia.com/embedded/community/reference-platforms
7. NVIDIA GPU Ventures Program
This is initiative created by NVIDIA to identify and support the best emerging companies worldwide to develop their businesses around GPU-based platforms.
If you’re already using GPUs and have interesting solution to important problem it may be something for you! Applications are accepted from fields as diverse as artificial intelligence and life sciences to automotive, robotics and virtual reality.
More info: http://www.nvidia.com/object/gpu_ventures_program.html
Portfolio of existing companies: http://www.nvidia.com/object/gpu-ventures-current-companies.html
Have promising project leveraging NVIDiA tech? Our suggestion is to show it off first in the challenge — nvidia.challengerocket.com to get your foot in the door.
One of the cool things you’re getting as a finalist is an opportunity to show off your project to CEOs, executives, and industry peers at the world’s biggest event for GPU innovation and AI conference, the GPU Technology Conference (GTC) in Silicon Valley.
That could be great way to get some serious exposure for your project and get in the NVIDIA ecosystem! (and as a finalist you’re getting full trip package as part of the award).
Tomasz Florczak
The views and opinions expressed in this article are those of the authors and do not necessarily reflect the official policy or position of NVIDIA.
About the author
Tomasz has been obsessed with technology for as long as he can remember — starting with the first computer codes written at the age of 7 (Basic and Pascal!). Some of his interests include AI, optimization algorithms, data analysis, urban traveling and discussing world geopolitics over a steaming cup of coffee!
Recipient of multiple brand awards (i.e. Coca-Cola, Al Jazeera, Samsung and Qualcomm). Occasional speaker at tech conferences with pretty impressive track-record of people not falling asleep :-) Co-founded ChallengeRocket.com platform and helped it to score 1st place in the European Startup Days Challenge. Now helping other brands to run successful engineer and software developer challenges & hackathons. Also volunteer assisting police and Ministry of Albania to fight human trafficking problem with digital technologies.
More:
https://challengerocket.com/pl/tomflo
https://pl.linkedin.com/in/tomaszflorczak
About ChallengeRocket
It’s a place where you can compete, show off your talent, win prizes, learn new skills and get discovered by one of the top global brands.
Founded by a group of technology geeks who has participated and won many international developer competitions. Now we are helping other brands to source innovation and reach out hidden tech talents with open challenges & hackathons. And we’re building community of thousands of talents looking out to be challenged.
Some of the most exciting projects we’re currently working on:
nvidia.challengerocket.com
accorhotels.challengerocket.com
We’re still a small team but we’re super happy to be noticed across the globe for the work we’re doing for our clients!
Psst! We’re building a new, top secret project right now. Please just don’t tell anyone. It will be about AI, predictive analytics, tech challenges and recruitment. We’d like you to be a part of the process. Subscribe to our newsletter and catch first updates from there (and get weekly news about new competitive-events we’re running on our platform). In case you’d like to get early access and become beta-tester let us know and we’ll count you in!
If you have any queries regarding this article or our organization, please write to us at office@challenge.com or reach the author directly at tomek@challengerocket.com.
