Book Image

Data Augmentation with Python

By : Duc Haba
Book Image

Data Augmentation with Python

By: Duc Haba

Overview of this book

Data is paramount in AI projects, especially for deep learning and generative AI, as forecasting accuracy relies on input datasets being robust. Acquiring additional data through traditional methods can be challenging, expensive, and impractical, and data augmentation offers an economical option to extend the dataset. The book teaches you over 20 geometric, photometric, and random erasing augmentation methods using seven real-world datasets for image classification and segmentation. You’ll also review eight image augmentation open source libraries, write object-oriented programming (OOP) wrapper functions in Python Notebooks, view color image augmentation effects, analyze safe levels and biases, as well as explore fun facts and take on fun challenges. As you advance, you’ll discover over 20 character and word techniques for text augmentation using two real-world datasets and excerpts from four classic books. The chapter on advanced text augmentation uses machine learning to extend the text dataset, such as Transformer, Word2vec, BERT, GPT-2, and others. While chapters on audio and tabular data have real-world data, open source libraries, amazing custom plots, and Python Notebook, along with fun facts and challenges. By the end of this book, you will be proficient in image, text, audio, and tabular data augmentation techniques.
Table of Contents (17 chapters)
Part 1: Data Augmentation
Part 2: Image Augmentation
Part 3: Text Augmentation
Part 4: Audio Data Augmentation
Part 5: Tabular Data Augmentation

Computational biases

Before we start, a fair warning is that you will not be learning how to write Python code to calculate a numeric score for computational bias in datasets. The primary focus of this chapter is to help you learn how to fetch real-world datasets from the Kaggle website and use observation to spot biases in data. There will be some coding to calculate the fairness or balance in the datasets.

For example, we will compute the word counts per record and the misspelled words in the text datasets.

You may think all biases are the same, but it helps to break them into three distinct categories. The bias categories’ differences can be subtle when first reading about data biases. One method to help distinguish the differences is to think about how you could remove or reduce the error in AI forecasting. For example, computational biases can be resolved by changing the datasets, while systemic biases can be fixed by changing the deployment and access strategy of...