AIMadeEasy

The Foundation for AI Futures
Imagine being able to extract 8x more insights from your data at lower cost and faster time-to-value. Join DDN, NVIDIA and other thought leaders for this important digital community event to learn how "AI made easy" can turn AI into business value.
Peter Burris
Given your processes and tooling for data preparation, how would high-performance infrastructure improve the speed and simplicity of your data ingest jobs. https://www.crowdcha...

Peter Burris
How does performance of technology translate into accuracy and quality of models and solutions? https://www.crowdcha...

Darrin Johnson
Fundamentally more data leads to better accuracy and quality of models.
Darrin Johnson
Being able to serve that data with low latency and high throughput is critical so clearly the performance of the technology is important.
George Vacek
The microscopy example given earlier is really telling. Performance yields the ability to get higher resolution images out of the instrument, whether for medical purposes or research. But also leads to real-time analysis as well which can be critical for insight
Kurt Kuckein
Real-time is just so essential for so many AI use cases, the I/O can't hold you back when decisions need to be made instantaneously, whether it is in retail, medical or government use cases
Yvonne Walker
I would love to hear what thoughts some of our viewers have!
Darrin Johnson
@DrV_DDN George you are exactly right. Going from 512 x 512 pixels to 4K images creates a critical performance challenge. Plus adding multi-spectral data it gets even more interesting.
sujinyan
The higher the performance the shorter the TAT. Gives the option of doing additional trials on the same hardware.
Peter Burris
What AI or deep learning use case are you most interested in? https://www.crowdcha...

George Vacek
Medical imaging is definitely critical
Darrin Johnson
Medical imaging to me is one of the most important fields for applying deep learning because of its direct impact on outcomes for patients.
Darrin Johnson
Obviously autonomous vehicles require AI and is probably one of the most data intensive cases out there.
Yvonne Walker
Genomic processing!
Brian Dowdy
@darrinpjohnson Totally agree Darrin... AI and the application of deep learning toward medical imaging will immediately impact us as a society and personally.
George Vacek
@darrinpjohnson The application to personalized medicine more broadly is also interesting
George Vacek
@YvonneW6463 There are impressive results for acceleration of genome analysis by Parabricks https://www.ddn.com/...
Kurt Kuckein
multi-disciplinary research into AI is fascinating to me, how do the different disciplines learn from each other?
Yvonne Walker
@darrinpjohnson makes a great point - data scientists just want the results, they don't want to be IT admins. This solution is making it happen!
George Vacek
Yes, easy drop in use for all the commonly used frameworks is key.
Darrin Johnson
Check out ngc.nvidia.com to be able to download containers for every major deep learning framework plus a number of HPC applications. Seconds to deploy and start gaining insight.
George Vacek
@darrinpjohnson Yes, exactly, that is fantastic!
Brian Dowdy
@darrinpjohnson Important plug for NGC Darrin! So many folks are interested but maybe weren't in the know of these nicely supported containers.

(edited)

Peter Burris
How mature is your adoption of AI? https://www.crowdcha...

Darrin Johnson
Working with customers I see a full range of maturity. Starting with people using GPU on the cloud or a couple GPUs in workstation to full production. However it is always key to think about managing the data.
Kurt Kuckein
We encounter such a range of customers in the AI adoption journey, from those just looking into it, to others with wide adoption - so many we are talking to now "just don't want to get left behind".
George Vacek
Not directly answering the question, but the easy scalability of the presented solution makes it applicable for those with relatively small amounts of data, but growing with them as they prove out their results and in turn gather more data.
Yvonne Walker
My thoughts on the game changer - DDN's ability to fully saturate those GPUs!
Darrin Johnson
Your exactly right...DDN's nearly 20 years of experience serving the HPC community brings both the experience and the technology critical to Deep Learning.
George Vacek
Nice example! Acceleration of microscopy image analysis not only speeds to real-time, but also improves the resolution overall. Incredible results from DGX-1 + AI200
Yvonne Walker
Agree! What an improvement!
Darrin Johnson
It is great to see how the combined solution makes a difference and this is only one of many examples.
Kurt Kuckein
@darrinpjohnson no need to compromise on capabilities when selecting and easy to implement and support AI solution
Darrin Johnson
@kkuckein your exactly right. Balanced solution architecture eliminates the bottlenecks.
Darrin Johnson
Throughput and low latency are key to maximizing Deep Learning performance.
George Vacek
NVIDIA GPUs can greatly accelerate the compute, but only if they can be kept fed with data.
Kurt Kuckein
Looking for information on AI use cases with DDN Storage and NVIDIA's DGX-1? Check 'em out here: http://www.ddn.com/?...
Yvonne Walker
I love how this WP talks about the AI enabled data center!
Darrin Johnson
Especially how it simplifies the solution for the customer. No guess work/experimentation needed to get performance.
Kurt Kuckein
How is storage a key element of deploying AI infrastructure? How does DDN maximize NVIDIA DGX-1 performance? Find out here: https://www.crowdcha...
[LIVE CHAT] The Foundation for AI Futures
We are live right now! Welcome to this important Digital Community Event. Come watch thought leaders from DDN, NVIDIA, and Wikibon discuss AI infrastructure.
Darrin Johnson
A key message - make it simple. The A3I solution does exactly that.