Join the data-centric revolution and master the concepts, techniques and algorithms underpinning the future of AI and ML development, using Python Key Features Get up and running with different data management strategy Learn how data-centric approach can help you get better Understand various facets of implementing data-centric approach in machine learning Book DescriptionThis book is for data science professionals wanting to understand what data-centricity is, its benefits over a model-centric approach and how to apply a best-practice data-centric approach to their work. This book is also for other data professionals and senior leaders wanting to explore the tools and techniques to improve data quality and how to create opportunities for “small data” ML/AI in their organisations. Data centricity is a new development within data science and learning materials and informational content on the topic is relatively limited. This book provides a rare end-to-end overview of data-centric ML, including hands-on application of technical and non-technical approaches to generating deeper and more accurate datasets.What you will learn Learn the principles of data centricity and how they apply to real-world scenarios Understand the impact of input data quality vs. model selection and tuning Know why subject-matter experts are critical to good model development Implement best-practice data cleaning, labelling and augmentation Learn common synthetic data generation techniques and their application Get well versed with applying these techniques using common Python packages Detect and correct bias in a dataset using best-practice techniques Appreciate the need for reliable, responsible and ethical ML/AI Who this book is forIt is expected that the reader has a good handle on the ML/AI process in general, and foundational knowledge of statistical methods and Python, if they want to complete the most technical parts of the book.

Book Information
ISBN 9781804618127
Author Jonas Christensen
Format Paperback
Imprint Packt Publishing Limited
Publisher Packt Publishing Limited