In this video, we'll take a look at IBM Cloud Paks. We’ll first learn the basics around Callbacks and then look at where they run. We'll take a look at each Cloud Pak in more detail and end with a demo showing how to use the Cloud Pak to train an AI model. Let's get started. What are IBM Cloud Paks? Cloud Paks are containerized software solutions built to run anywhere? I like to think of it as a bundle of continuous software. The goal of the Cloud Pak is to make container management and application modernization easier for an organization. Cloud Paks come in a variety of use cases, which we'll explore later. For now, let's look at the values of Cloud Paks. Cloud Paks have a modular architecture, meaning you can pick and choose which software you want to deploy. Cloud Paks put AI to work. They operationalize AI throughout your business. Cloud Paks are built on OpenShift, which allows them to run anywhere. So, what do we mean by anywhere? Cloud Paks can be run on any platform you choose. You can install them on IBM Cloud. You can install them on premise. Use your own hardware or with the Cloud Pak system, or you can install them on any cloud by first provisioning OpenShift and then installing Cloud Paks on top. The following picture is a representation of how different Cloud Paks are available. The six Cloud Paks that we will cover are applications, data, integration, automation, security, and multi-cloud management. The Cloud Pak for applications has tools to help you modernize existing applications and build new cloud native ones. The four main capabilities are: Accelerator' for cloud native development, which bring together open source technologies and put them in a microservices based framework. IBM modernization and developer tools. Modernization guidance gives you a plan to start strategically updating your applications. A Java EE platform is a collection of Java APIs that help you write secure, flexible server side applications. Mobile app development tools for building apps for mobile, wearables, conversation, web or progressive web apps. Cloud Pak for data has the following features built in: It has a single platform that integrates data management, data governance, and analysis. It has databases so you can spin up your favorite IBM or open source database. It has data governance built in, like automated discovery and classification of data, and masking of sensitive data. It also has data virtualization so you can query easily across multiple sources, on cloud or on premises. It has IBM Cloud AI services, such as Watson Assistant or Discovery. And it has AI model lifecycle tools, such as Jupyter or RStudio to create notebook, serve them with Watson Machine Learning, or automate the whole process with Auto AI Cloud Pak from Multi Cloud Management is an IT management platform designed to provide full visibility and control wherever your workloads run. Using the dashboard, you can monitor application lifecycle management so you can deploy and move application across clouds. It provides cloud protection and compliance with automated policy enforcement and compliance testing. It has built in SRE tooling with AI OPS to use event correlation and machine learning to improve operational efficiency and readiness. It also has many add-on capabilities from IBM partners such as Turbonomic, Sysdig, Humio, and Hazelcast.. Cloud Pak for integration is a complete set of integration capabilities to efficiently connect your applications and data wherever they live. You can use API connect, app connect, MQ and event streams and you have the Aspera high speed data transfer so you can move data of any size around the world at maximum speed. IBM Cloud Pak for security is a platform that helps you uncover hidden threats. Make more informed risk-based decisions and prioritize your teams' time. It has core platform services such as threat, intelligence, insights, and data Explorer, and has integrations with existing tools and data such as Qradar and Splunk. The IBM Cloud Pak for automation provides applications in core areas where automation provides benefits, content, workflow, decisions, and capture cloud fact for automation. Provides low code consumable tools, API's and application connectors that make it easy to consume content in business applications. You can also automate your end to end workflow with IBM business automation workflow and you can use IBM operational decision manager to automate the implementation of business policies in your organization. So, now let's dive a little bit deeper and show you an example of IBM Cloud Pak for data in action. So, here we're logging in with our username and password into our Cloud Pak port data instance. We click sign in and here we're greeted with the overview page. Here we can see our projects or instances and if we go to the top right, we can see our services catalog. Here we have all the AI services, such as assistant discovery, knowledge studio, and the one that we're going to use today is Watson Studio, and we see it's enabled. And now we can go back to our overview page and we can click on projects to create a new project and we’ll create an analytics project and then we'll create an empty project. But you can also create a project from a file, but now we'll just do an empty one and we'll call it Horea demo. We’ll click create. And now you can see that we're into Watson studio here and in our Watson Studio instance we can click on auto AI experiment. That's going to give us a new experiment which will automatically train AI model. We’ll call it demo auto AI and we'll give it a compute configuration and now we have to give it a data set. So, this is an insurance data set from Kaggle. Using age, sex, BMI, number of children, smokers, whatever region you're in and then it's trying to predict expenses. So, we're predicting how much is an individual person going to pay for their insurance expense, is are they going to pay 16,000, 1,700? 4,000, 21,000, and it's all based on if they're smoker or not, their BMI, their age and other factors. So, we're going to use that insurance file that I showed you earlier, and we're going to feed that into Auto AI. Auto AI is going to automatically detect what type of data that column is and we're going to predict expenses And you can see it's a decimal type. So it's going to be a continuous value from zero to Infinity, and we're going to predict based on all of these factors and all this training data, once we give it new data, you know what... is my insurance going to be? My insurance charge going to be? You can see here in auto AI we can see our different pipelines being created. Here we can see the model itself. We can rank by different metrics such as R-squared, root mean, squared error. These are all different ways to see how well the model is doing once you feed it new data and test it with new data. You can see the algorithm too, using random forest regressor or gradient boosting regressor. And here we can save it as a model. So, now we've saved it as a model and already deployed it. And once we've deployed our model, we can go ahead and use it via rest API. So, you'll see this is our endpoint, the Zen1-CPD. So you can copy that and use a rest API call, and you can use Curl, Java, JavaScript. And we have these code snippets ready for you to show you how to make a post request. So, here's the input data that we saw based on what we saw earlier. Age, sex, BMI, children... So, I'll fill in this details. Zero, smokers no, region of Southwest. If I click predict, it's going to use this train model that we've deployed and it's going to come up with a prediction. So, we have 4,000. If we say yes for smoker, the value should be much larger because if you're a smoker, you're going to pay more for your insurance. So you can see the value is much larger at 16,000 instead of 4,000 before keeping all other data points equal. So, this was an example of using Watson studio within Cloud Pak for data and showing you the interface and how easy it is to use auto AI to train a model. And now, to summarize, there are six Cloud Paks. Cloud pak for application, cloud pak for data, cloud pak for integration, cloud pack for automation, cloud pak for multi cloud and cloud pak for security. Cloud Paks are a collection of software that is modular and you can choose to install all components or just a few. And Cloud Paks are based on OpenShift, meaning that they can be run anywhere, IBM Cloud on Prem or any cloud.