Book Image

Artificial Intelligence Business: How you can profit from AI

By : Przemek Chojecki
Book Image

Artificial Intelligence Business: How you can profit from AI

By: Przemek Chojecki

Overview of this book

We’re living in revolutionary times. Artificial intelligence is changing how the world operates and it determines how smooth certain processes are. For instance, when you go on a holiday, multiple services allow you to find the most convenient flights and the best hotels, you get personalized suggestions on what you might want to see, and you go to the airport via one of the ride-sharing apps. At each of these steps, AI algorithms are at work for your convenience. This book will guide you through everything, from what AI is to how it influences our economy and society. The book starts with an introduction to artificial intelligence and machine learning, and explains the importance of AI in the modern world. You’ll explore how start-ups make key decisions with AI and how AI plays a major role in boosting businesses. Next, you’ll find out how media companies use image generation techniques to create engaging content. As you progress, you’ll explore how text generation and AI chatbot models simplify our daily lives. Toward the end, you’ll understand the importance of AI in the education and healthcare sectors, and realize the risks associated with AI and how we can leverage AI effectively to help us in the future. By the end of this book, you’ll have learned how machine learning works and have a solid understanding of the recent business applications of AI.
Table of Contents (10 chapters)

One-shot learning and transfer learning

If you dont have sufficient data to train deep learning algorithms, there are three ways to work around it: generate synthetic data, scrape/buy data from external sources or develop AI models that work well with small data.

Deep learning is very data-hungry — models are trained on huge sets of labeled data, e.g. millions of tagged animal images — and large amounts of labeled data are not available for specific applications. In such cases, training an AI model from scratch is often difficult, if not impossible.

As weve mentioned, one potential solution is to enlarge real datasets with synthetic data by generating more examples. This has been successfully used in autonomous driving, where autonomous vehicles drive millions of miles in photorealistic simulated environments that recreate situations like snowstorms and unusual pedestrian behavior and where acquiring real-world data is hard.

Similarly, researchers...