In this introduction, we define the term bias as it is used in machine
learning systems. We motivate the importance of automated methods for
evaluating and selecting biases using a framework of bias selection a
s search in bias and meta-bias spaces. Recent research in the field of
machine learning bias is summarized.