Hi folks, Chris Gonzo Gonde here again for another newsworthy minute with NetApp. Today we're going to be talking about artificial intelligence, machine learning, and deep learning. So AI is changing the game in every industry from medical imaging to fraud detection, smart cities, and driverless cars. But AI success is easier said than done. Training artificial intelligence algorithms requires data to flow like water through a fire hose. So, we've partnered with Nvidia and Fujitsu to give you more AI with more data and remove those AI infrastructure obstacles and bottlenecks to optimize performance from edge to core to cloud. So, I'm very excited to introduce our guests today. We have Chris Bested from our very own uh NetApp team and we also have our special guest Joel Lopez uh from Fujitsu. Welcome Chris and Joel. >> Hello. Hi, Chris. Good to be here. >> Awesome. So look, um, folks, I'll get you to tell us a bit about yourselves and your roles. Um, Chris or Bestie, as I, uh, affectionately refer to you, I'll start with you. >> I'll respond to anything. Hey, Chris. Um, my role is to create value for NetApp and our partners by essentially developing solutions that leverage NetApp's extensive portfolio of data fabric capabilities. Um, see, I've always been very passionate about a co-creation approach as a way of capturing people's imaginations to achieve business goals, and I'm very excited to be um have the opportunity to tell everyone about this great initiative that we've developed with Fujitsu.>> Awesome. All right, Joel, over to you, mate. Let us know a bit about yourself and your role at Fujitsu. >> It's Joel here. Thanks for that, Chris. um mainly responsible for delivering solutions to our customers um AI related solutions, high performance computing and also cloud computing um platforms and um I think this is a good opportunity to talk about more on what really we are doing. >> Fantastic. Okay. Well, look, I think um Joel, you're probably going to be the perfect person to start with. My first question,and this is something that I've kind of thought about since I heard the terminologies, right? Artificial intelligence, machine learning. These seem to be overused and largely misunderstood terminologies. So, can you tell us a little bit about well, our offering we call onap AI and its relevance to AI, machine learning, and deep learning? >> Thanks for that, uh, thanks for that Chris. Well, on AI is an intelligent platform and architecture. I would say it's a modern platforms for modern workloads. But before that, before I answer your question directly, let's have a quick look and understand what really AI is. Right? AI is nothing new. It's been um since 50s u AI is already being used for games or even in sci-fi movies. But uh they are restricted or limited when it comes to application back then, right? And if you guys remember in the '90s uh if you see um an email spam on your um computer that is actually a product of machine learning. So machine learning is a part of AI and that email spam that um we have invented or we have seen is a machine learning techniques, right? making it intelligent to identify that is email is good or email is bad. But within the machine learning there is what we call a subset whatis called deep learning a lot deeper and in deep learning has been popularized or being uh more um used by data scientists or researchers. Uh as of today there are over three million um researchers or data scientists using this data or what we call deep learning techniques in order to extract more information from large data sets. But when we say large data sets where all these data came from? Well 10 years ago when we heard about data explosions coming from everywhere right mobile phones uh coming from um your smartwatches etc. I think for the last 10 years the world has already um accumulated huge amount of data for everyone to analyze right even our customers are now taking advantage of this data to explain and to generate useful information from this. So data connecting to deep learning because like I said deep learning is a technique and a subset of machine learning to use deeper algorithm so that I can extract more information into this big data and then in that way I can use that for my research. So deep learning is reallyuh popular now and people are or researchers are taking advantage of it to get a better and efficient uh answer to the questions. >> Soyou said a couple of key things there like um provisioning the data to the deep learning uh mechanisms and things like that. So I guess my question is what do we need before we can really leverage the machine learning technologies that we are delivering? What else do we need besides the data? That's very good Chris and um I think I missed one of the information important information earlier in the earlier question about the value of on top AI right>> so >> so >> so the value of one type A is reallyimportant we have researchers they have techniquesthere there's a big data into it so what do you think in mind is needed to analyze and crunch the data using that what we call convolutional network that's the term that researchers are using right they're using convolutional network deeper algorithm to analyze that what do we need a high performing AI optim optimized platform >> today GPU based computing is the most dominant when it comes to analyzing this data but we have >> Nvidia partnership >> correct but what do you need aside from the GPU powerful GPU you need a high performing storage that works together with a GPU to deliver that capab capability and to make it very quick, very fast when it comes to getting the answer and that is why the on top AI architecture is a perfect fit for that. Netup in Nvidia has developed this uh architecture and solution todeliver that solution in this high demanding workload because if you want a high-speed answer, what do you need? You need high performing solution >> that data that flows like water from a fire hose like we were talking about earlier. [laughter] earlier. [laughter] earlier. [laughter] >> That's correct. And now to answer your second question about what do uh our researchers or scientists need, we've been Chris and I have been working with um a lot of um data scientists for quite some times andwe were also learning along the way and most of the learnings that we have found is that um they need to define the right strategy on their projects. you know most of the time they will do the trial and they will um experiment uh thetype of the platforms or frameworks and tools that they need. Today there are many modern tools and frameworks that they can use to run their AI projects from experimentation, prototyping to production,>> right? So at first they need to define the strategy and the frameworks and tools and number two is they need to know how much amount of data they have. The more data the better success they can have from prototyping to production. And once they have this data in their uh in their framework and tools, the next thing they will do is to find how much computational resource they need andthese are the things that are reallyexciting and interesting since we started working with uh many uh researchers and scientists. >> Okay, brilliant. Well, it sounds like Fujitsu is in a really unique position to take our uh combined technology and go to market strategy around AI. Um so Chris I'll switch to you for a second and tell me um I understand that we've developed an AI test drive capability. Um can you tell us a little bit more about this and what the benefits are for our customers and partners? Yeah, sure. Yeah, sure. Yeah, sure. Chris Luke, I think you know the AI test drive is a platform essentially that we've um pulled together with combines the latest data management platform from NetApp of course the GPUs and software from Nvidia and AI and expertise and HPC pedigree that Fujitsu bring to the table. So it's a combination of all those three and as Joel described you know it's built on the industry recognized on AI reference architecture but it's a essentially a demonstration of how this reference architecture can remove infrastructure and data management obstacles and bottlenecks that data scientists face today. So we've worked very closely with Fujitsu to bring this to market using this reference architecture and um you know thepower of it comes when we can extend give them access to their data across any of the three major realms where the data may be to get today whether it be at the edge um inside their DC or maybe in existing cloud workloads. >> Okay. So it sounds to me like this culmination has created what uh we refer to as the AI trusted platform. Um can you tell us a little bit about what is the AI trusted platform and how are we taking this to market with Nvidia and Fujitsu?Yeah, sure. You know, um I would say this is not a simple question to answer because no one just goes to market to buy some AI, right? The major issue with adoption of AI is that it often fails to live up to the hype that surrounds it. Um so a natural indicator that we've observed of this is the immense amount of pressure that data scientists are under to deliver an outcome. And data scientists typically will be forced to adopt ajack-of- all trades approach and become knowledgeable in many unrelated fields >> to that of actual data science areas such as the data movement and preparation uh security compliance um ethics communication they even need to be you know business acumen to manage perception. So um you know there's a lot riding on the shoulders of data scientists to deliver on that outcome and I guess the AI test drive is an exercise ultimately in co-creation but it's more so than just the three parties that um it puts their clients at goals at the center um and it provides a way which we can bring a purpose-built tangible tools and processes that can help data scientists focus on the job of data science and net and video have created that robust validated the validated architecture which we call an ONAP AI. The challenge was how do we connect the value of this platform that it and that it provides to the people that will benefit from it most. And that's ultimately what the AI test drive is all about. It's about putting the customer's objectives first. >> Awesome. All right. Well, look, um, uh, Joel, I'm going to finish with you. Um any advice for aspiring data scientists or partners out there who are looking to get started with our AI test drive capability?>> Thanks for that Chris. uh look the most important thing that they need to have will be number one data like I mentioned earlier and uh the frameworks and tools like I said earlier modern tools are available now for them to utilize it and what they can um experience in the test drive are even additional modern tools from data ingestion data manipulation uh data security, governance and compliance and the most important is the most powerful computational resource they can use to train or benchmark their data and to validate their algorithms giving them an opportunity in a very quick way to validate um their prototyping and maybe helping them work through up to production right andthese are all available I I'm keep on using the words modern tools because andhigh performing workloads because that is how the AI test drive based ontop AI has been developed and delivered today to help researchers and data scientists. >> Fantastic. Joel, Chris, thanks so much for joining me on this uh AI focused uh newsworthy minute. Uh real pleasure to have you both uh on board and talking about um not just thepartnership between Fujitsu uh NetApp and Nvidia but also this um ready to go AI test drive capability that uh we're doing together. So um uh thanks again for your time. Looking forward to catch up catching up with you again in the future to hear about how we've applied this to uhcertain projects out there with data scientists and uh until then thanks again.>> Thank you. Thanks for your time Chris.