Employing Machine Learning in AI
IN THIS CHAPTER
Using the tools of different tribes when learning from data
Discovering how probability benefits AI
Guessing using Naïve Bayes and Bayesian Networks
Partitioning data into branches and leaves by decision trees
Learning has been an important part of AI since the beginning because AI can mimic a human-like level of intelligence. Reaching a level of mimicry that effectively resembles learning took a long time and a variety of approaches. Today, machine learning can boast a quasi-human level of learning in specific tasks, such as image classification or sound processing, and it’s striving to reach a similar level of learning in many other tasks.
Machine learning isn’t completely automated. You can’t tell a computer to read a book and expect it to understand anything. Automation implies that computers can learn how to program themselves to perform tasks instead of waiting for humans to program them. Currently, automation requires large amounts of human-selected data as well as data analysis and training (again, under human supervision). ...