Sri Ambati at Ai4 Conference 2021

[Music] so [Music] welcome to the third and final day of ai4 2021 it's been quite a ride and we have another amazing lineup of speakers to learn from today along with equally notable attendees to network with today is your last chance so reach out and make connections with all of those you haven't met yet before we get started today here is a 90 second commercial from our sponsor h2o [Music] h2o ai hybrid cloud is an end-to-end platform that enables organizations to rapidly make world-class ai models and applications for virtually any use case building ai models used to take large teams of expert data scientists months or even years with h2o ai hybrid cloud an individual data engineer developer or data scientist can make world-class ai models and applications many times faster rapidly putting ai into the hands of business users [Music] h2oai hybrid cloud brings automation capabilities across the entire data science lifecycle including connecting to and repairing data building and explaining models and deploying and operating them the platform also makes it easy to publish and share ai applications across an entire organization h2o ai hybrid cloud runs on kubernetes allowing customers to operate on any cloud or on-premise infrastructure if you're looking for a solution that enables you to put ai into the hands of every employee get started with h2o ai hybrid cloud one platform thousands of use cases [Music] first up we have sri ambati shree is the founder and ceo of h2o.ai where he strives to make artificial intelligence easily accessible and democratize ai for everyone sri's vision is to make your company an ai company which he has successfully applied to businesses and organizations in a wide range of industries from finance to health care to retail and manufacturing sri continues to help companies maximize the impact of their data on business results and improve day-to-day life for people around the world through the h2o.ai platform sri founded h2oai in november 2011. prior to that shriek co-founded platfora he also worked as the director of engineering at enterprise data stacks he also worked as an engineer for us systems and right order inc sri got his bachelor of technology degree from srkr engineering college and his master of science mathematics and computer science degree from university of memphis please join me in welcoming sri to the stage [Music] all right please welcome shree to our virtual stage thank you patrick h2o is here to democratize ai we make a.i ubiquitous and when we think about democratizing ai we think about faster cheaper easier every everyone should have ai in there in their hands time is the only non-renewable resource and speed is what matters in everything that we do and a lot of our customers have adopted ai over the last decade some of our strongest customers have been creators of ai themselves and makers of ai applications for their business users at the heart of some of these customers have been their transformation that they brought not just to their cultures as companies but also making their themselves fast learning organizations one of the top supporters of this conference capital one wells fargo folks who will be hearing after this have been adopting and embracing h2o and ai over the years we are at the heart of every payment system out there at the core risk modeling and and most of the biggest banks in the world alike apple pay paypal and wells fargo some of the core risk models in all the banking systems need that transparency and the trust that one can depend on att said a billion dollars last year on fraud prevention for the iphones we've seen uh fraud prevention be a big landing use case for us at several of our customers but beyond that subprime credit scoring to actually um cyber and anti-money laundering as well as know your customers some of our customers have gone from being number seven in rental insurance in north america to number one and ai has powered those transformations for some of them ups for example re routed every one of their uh packages in the pandemic which was a surge of volume for them and they used h2o ai for that and we found that every mile saved for some of these delivery systems is millions of dollars saved walgreens assessed flu shots in every corner of every zip code how many people will be hesitant for a vaccine versus how many people are open to taking one progressive insurance uses h2o to look at which bend in highway on cleveland it's going to lead to more accidents and make their customers safer critics fees h2o have built applications some of which we will demonstrate today on prevent on changing contracts to help during the libor transitions our customers like paypal fight collusion fraud between buyers and sellers e-business has taken a new level of um of volumes but also with volumes came a lot of risk our customers like discover are using us at the heart of their risk modeling and making ai more explainable to their regulators mastercard powers billions of transactions today through uh h2ai fraud prevention systems we're working with marcus on goldman sachs at goldman sachs and also several of our customers in the data business supply chain has been truly impacted during panamic and our customers have been at the heart of adopting distribution centers in the globally impacted changes in demand needless to say customers are at the heart of our movement and they have been the true supporters of our business at capital one we went from one use case in the payment system to several use cases in the consumer side auto finance kyc cyber security aml document classification to subprime credit scoring and transaction and application fraud prevention we're super excited for a partnership with some of our largest retail banks in the world every company it needs to be an ai company and we think that this vision has been as coming is coming to fruition one of the earliest dot ai domains when we started and democratization of ai meant that making ml making ml not just fast cheaper and easier to use but also the hands of every customer and every data scientist but ai is not just a technology transformation it needs incredible teamwork and that teamwork has made this stream work our customers a lot of them are leading um awares of data but they're at the heart of their mission is success of their customers and it starts with domain great algorithms still need incredible partnership with business users and designers to make that impact that's needed to make decisions happen at the edge of every one of these different professions is an incredible amount of of velocity but also conflict and conflict is not just a bug it's a feature it in that conflict is where we've gotten ourselves to innovate and build tool chains and products that address the communication between these communicative communities within your own organization data is a team sport and explaining those models how they work how the different dimensions the data scientist is 100 percent passionate about data well an executive is more interested in storytelling and hopefully we will all be only telling more stories as we as we go on to be up to up and being more human it is this explainable ai that sought of making ai accessible to the business users and the end customer community our innovation started with ml and making ml really accessible to the world's largest data science group from ml we transformed ourselves to auto ml and on auto mld our products like driverless ai are the leading platforms for customers to build reusable components recipes that allows ordinary data scientists or new data scientists to become as powerful or as experts as the capital grant masters our ai middleware and apps and app stores are going to further democratize ai and make ai as ubiquitous as software or hardware or technology in your in your company it is these app stores that are going to power the future trillion dollar companies with the ai and applications are going to make ai accessible to business users so our distinct vision is to make ai subliminal to your applications subliminal to your corporation and your organizations an example here is we're taking all the reusable components design patterns needed for making ai first applications in a bank and then making those app stores actually accessible to our customers some of these applications are even open source we built an ai low code environment that allows you to build these applications and rapidly prototype ideas and put it in front of your business users for more um for feedback for feedback is actually the ultimate goal of all ai our ai app stores are being co-created with our customers co-created with our community and with our ai for good mindset with and and volunteers from across different uh yeah for good initiatives worldwide we think that a centralized marketplace is very much possible not just for us but for our customers and making these ai app stores accessible to our customers because their customers becomes an even more of a mission over time we expect ai apps to be the hands of not just our customers but their customers and the suppliers of their customers we think that this is the beginning of a very long epoch in building applications and the life cycle management of these models and these applications becomes the heart of how we deliver ai ai first applications are needed for business expansion they are at the heart of every um collaborative movement in your organization and with democratized access and the freedom to innovate and just purpose and data can power the dream of any ai entrepreneur out there over time we expect the data to transform itself to models into models with the powerful ai middleware and ai cloud we're able to deploy ai apps and deliver them to several of your customers just some of the examples of our customers are co-creating and innovating with us in healthcare finance and insurance software is leading the world and ai is using software and hardware to be precise and data is going to be the new code from mere changes in your data tables you can produce logic with ai that means today the companies with the largest data ecosystems are going to be the leaders of the future so you discover your competition it's not just the next vertical neighbor it's actually your largest tech giants that are out to take advantage of the data and the cloud and deliver even more powerful customer personalized customer experiences for your customers h2o is here to make you a strong ai superstar superpower yourself and it's going to take a lot a lot of data ecosystems and data products for us to get there the defensive brand and community with open source and data is at the heart of our movement and open source is about freedom not for you our open source custom [Music] products are used by thousands of customers and millions of open source data scientists we think that open source is going to breed such rich ecosystem we are here to raise not a force not just a tree making ecosystems will be the heart of every ai company and we think that all our customers are beginning to realize that and the app stores deliver rich ai ecosystems ai for good is a moment that we are strongly have been embodying that since the making of our company whether to democratize health which is how we got started or our investments in responsible ai and fairness making a.i more transparent has been at the heart of our movement to prevent ethical bias or or over the last hundred editions of credit have embodied by us in them how do we find that with much more powerful algorithms and tools that can elevate the voice of the unheard but more closely ai to protect wildfire for wildlife and pets we had to fight wildfires and model wildfire behaviors to prevent loss of life for ai to manage pandemic supply chain have been at the heart of our more recent co-creations and collaborations with the community we are we are very excited to encourage partnerships in the audience for folks who want to volunteer with us our customers our communities our open source users as well as all of ai4 to come come forth and partner with us in this mission to make the world better with ai as a call to action in the summary you're here to make ai hybrid cloud easy um and make it easy to make your own ai operate innovate and monetize with ai a call to action is to go to cloud.ai you can try this today let me show a few demos that demonstrate where we are going with our ai here is um some of the rich set of app stores that we built with our customers many co-created to build applications of course one needs a low code environment to build apps this is an app we built with our customer wells fargo what this is doing is essentially allowing customers to open up deep machine learning models for locally linear interpretation interpretable models are the heart of understanding how ai is working to prevent and fight bias but also to make sure you're not exposing yourself in front of regulators with the wrong kind of models this is an nlt model that's able to understand why a very positive sentiment is coming from this um from these um reviews of your products if you're a customer or a restaurant owner you want to understand why a particular model why a particular model is predicting a positive experience for your customer the entire end-to-end of this application was built co-created with our customers in this case wells fargo you'll hear from august later today and the idea is to be able to not just um not just be focused on the model building but also taking those complex sophisticated models and explaining them to your customers to your business users locally linear models are opened up that incredible ability to kind of go beyond just the deep neural network methods to actually going down and explaining why the model is behaving the way it's behaving we take another application and model models come at us at several forms and factors and our automatic machine learning is able to expose not just the model's interpretation but also you're able to go beyond that and understand and derive rules business rules from your models this is an application that is built out of on top of our automatic machine learning and is now able to derive a business rule from a cluster on top of your sharpening variables sharply as many of you know is is a game georgic method to understand and explain the why models behaving the way it's behaving and here you can essentially look at the business meaning behind a cluster of on top of a cluster that on after of the of the rules the idea is that take this and then start building and playing with it and understanding why not just the how our vision to make ai even more democratized goes beyond just trying to build the models to then take the models of production and reuse components that help you as a customer to thrive it's an application you're co-creating with credit space the idea that libor is going away is very common and there's lots of millions of contracts that need to be redone because of the changes that are happening in libor so replacing processing these contracts would take a lot of human effort and time and what we're doing is reducing the labor intensive part of that and allowing um humans to really focus on ones that are the more important ambiguous part of that decision edition lifecycle so what this application is is demonstrating is how to use birth models and to process contracts and then pull out entities where libraries and then try and replace and reduce the problem of half a million documents to just 500 entities that need to be changed again a reusable component that can that has the underpinnings of how to go make perhaps even more smarter some of our customers in manufacturing for example they take a look at the entire factory floor and they won't understand why an engine failed and instead of looking at the how which is kind of a classic monitoring and the modeling behind they're able to go beyond the hub to understand the top five factors that the engine is being predicted to fail the driverless automatic mission learning that we have is predicting this particular engine to fail and it's highlighting the why the top sensor 11 is off the max and since the four is of the max and those factors are what show up as the real things behind our vision for our customers is to kind of start going beyond just the simple to the double clicking and understanding why and how we think this this is just the tip of the iceberg so many different use cases that need to be automated and reused and make make models model building much more simpler and easier model deployment becomes a very important aspect of how do you take models to the next phase the life cycle of your model taking the model such as a forecasting model and understanding how it is changing both back testing it over course of time and then building adversarial similarity and and drift detection across the data so you know when to retrain the model and that ability to retrain model in reaction to changing data that's kind of where we think the end-to-end life cycle is today in the lobs is a separate aspect of building models and deploying models we think it needs to be integrated into how one builds models and that's kind of at the heart of our ability to look at models and explain them and understand how how models can be interpreted and then then truly democrat and truly test them with adversarial testing and take it to production this we think that is going to be explainability will be at the heart of building all models and taking into production in to suffice to say a lot of these models are built on the ai engines behind the scenes ai engines obviously which are familiar to our customers automatic machine learning to driverless ai ml from h2o open source and of course you can plug in several different data engines to truly give you the end-to-end um vibrancy that is needed to take your models to production models model monitoring is at the heart of the other pieces that you want to understand how your models are behaving on a daily basis and when you launch a model from our structure you have kubernetes which makes it model the deployment cloud agnostic you can take your models on any cloud and deploy it at scale we would love for you to give your feedback on how your experience of cloud.h.a.i.s give it a try and let us know how you transform your organization you think you think that making your own ai on any cloud is the is the road for transformation in making an organization a fast learning organization make your own ai app source with h2o you can try this today at cloud dot hti thank you thank you sri

Channel: H2O.ai

Views: 170

Rating: 5 out of 5

Keywords:

Id: G0zNQ0rsD28

Length: 24min 43sec (1483 seconds)

Published: Sun Sep 05 2021