AI on Google Cloud
1. AI on Google Cloud
SPEAKER: Inspired by the Coffee on Wheels use case? Let's explore how Google can help you turn your own AI-powered ideas into reality. So why Google? First, Google has a long history of leveraging AI to power its products, from Google Search to popular tools like Google Maps and Workspace. Google is eager to share its experience to empower you, whether an individual or an organization, in realizing your AI ambitions. Second, Google leads an AI in ML innovations, particularly with recent generative AI breakthroughs. Technologies and products like Gemini, Vertex AI, and NotebookLM exemplify Google's commitment to delivering powerful AI services for your projects. Third, Google believes in responsible AI. Google wants to collaborate with you to foster bold innovation, responsible development and deployment, and collaborative progress. Explore Google's responsible AI principles in the reading to learn more. And while this all sounds fantastic, you may be wondering, what is the current landscape of AI problems? To simplify, let's categorize the problems. Recall the Coffee on Wheels use case presented earlier in this module. What problems could AI solve for them? One type of problem is prediction, exemplified by sales forecasting and route optimization through traffic prediction. This is known as predictive AI. The other type is creation, such as generating customer responses and automating marketing campaigns. This is called generative AI. Let's delve deeper into predictive and generative AI, comparing them and determining their optimal applications. So what are predictive AI and generative AI? Predictive AI, also known as traditional or discriminative AI, uses existing data to classify information or predict future outcomes based on historical patterns. It excels at learning from what's already there to make informed decisions, much like using tried-and-true methods to and forecast. Generative AI, on the other hand, expands these capabilities to create summaries, uncover complex correlations, or generate new content. This includes text, images, or videos that mirror the style and patterns within the training data. It doesn't just analyze, it creates. To put it simply, predictive AI analyzes and predicts. Generative AI creates new content and takes action. Now, when you should use them is based on your use case. Here's a simplified decision tree. For forecasting and predictions, a predictive AI model might be your go-to. And for multimodal content generation-- text, image, video, and automation-- generative AI is ideal. However, there's no clear line between these two. And sometimes, the best approach is to use both. You can use the output from a predictive AI model as part of the prompt for a generative AI model. For example, use predictive AI to forecast customer churn. Then use generative AI to power a chatbot that helps your sales team explore these predictions. Or use predictive AI to identify customer segments. Then use generative AI to create personalized marketing content for each segment. By prioritizing business outcomes and user needs, you can maximize the benefits of both types of AI. You can think of the Google Cloud infrastructure in terms of three layers. It all starts with a robust AI infrastructure featuring advanced compute, network, and storage technologies. This foundational layer provides the solid ground for building your AI realm. Next, we move up to the development layer, where the real magic happens. Google's Vertex AI, an end-to-end AI development platform, guides you from design to deployment. Powering Vertex AI are Google's foundational models like Gemini and streamlined deployment pipelines. Vertex AI also seamlessly integrates with data products like BigQuery, ensuring a smooth journey from data to AI. This AI development layer is truly the ultimate playground for developers, engineers, and data scientists. And what if you're not a technical professional? At the top layer of applications and solutions, Google provides out-of-the box options for business users and analysts to rapidly prototype their ideas. So how does this course help you comprehend Google's AI architecture? You'll delve into AI infrastructure and the data tools in this module. Both lay the foundation for AI development and applications. Module 2 kicks off your AI project journey by focusing on building gen AI. You'll learn about gen AI foundational models and the tools needed to build AI projects at both development and application levels. And don't worry. We haven't forgotten about predictive AI. In module 3, you explore different options to train an ML model. And then, in module 4, you build an ML model end to end, from data preparation to model training and deployment. Excited? You'll learn about AI infrastructure soon.2. Let's practice!
Create Your Free Account
or
By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.