From Scripts to Skills: Learning-Based Control

Predefined scripts break the moment the world deviates from expectation. From Scripts to Skills: Learning-Based Control examines how robots move beyond hard-coded instructions and acquire adaptable, reusable skills through learning.
This book is written from a control and intelligence perspective. It focuses on how robots learn behaviors rather than execute commands—using data, feedback, and experience to improve performance over time. Readers are introduced to learning-based control as a shift in philosophy: from specifying every action to shaping how decisions are made.
Rather than presenting learning as magic, the book explores constraints—sample efficiency, safety during learning, generalization, and stability. It explains why combining classical control with learning produces systems that are both reliable and adaptable.
This book digs into:
• Why scripted control fails in dynamic environments
• How skills differ from sequences of commands
• How learning and control are combined safely
• Why feedback shapes behavior more than rules
• How robots transfer skills across tasks
This book is for roboticists, control engineers, and researchers who build systems that must adapt, not just execute. If your robots need to handle uncertainty without constant reprogramming, this book explains how skills replace scripts.