Hands-On Explainable AI (XAI) with Python: Interpret, visualize, explain, and integrate reliable AI for fair, secure, and trustworthy AI apps

Machine-learning predictions are easier to use when people can understand how a model arrived at them. Hands-On Explainable AI (XAI) with Python explores that challenge through practical examples, showing how explanation and visualization techniques can help make model behavior more legible—and raise important questions about bias, fairness, and responsible use.

Denis Rothman takes readers from core XAI concepts into Python-based applications, including examples involving medical diagnosis, self-driving cars, data exploration, and sentiment analysis. The focus is hands-on: examine models and data, interpret results, and consider how explanations fit into the wider machine-learning workflow.

Opening the black box with practical XAI

The book introduces explainable AI through both interpretable, “white box” models and methods for examining more complex systems. Early examples use approaches such as K-nearest neighbors and decision trees to make model decisions easier to inspect. The discussion also connects explanation to high-stakes settings, including medical diagnosis and autonomous driving.

Explore data, bias, and model behavior

Tools such as Facets support the exploration and visualization of datasets, while chapters on AI bias and ethics consider how data and design choices can affect decisions. These examples encourage readers to look beyond a model’s final prediction and examine the information and assumptions behind it.

Compare explanation methods

Rothman introduces a range of techniques, including SHAP, LIME, counterfactual and contrastive explanations, and Anchors. Examples across different data types help show how explanation methods can answer different questions about a model’s output. The book also explores the What-If Tool and cognitive approaches to explainable AI.

Who may find it useful?

This book is suited to readers with an interest in Python and applied machine learning who want to explore model interpretability in practice. Its examples may also interest data scientists and AI practitioners considering how to communicate model behavior and assess fairness in AI applications.

For readers building or evaluating machine-learning systems, this is a practical tour through the tools and questions that make AI explanations worth examining.