Inductive bias can be understood as an assumption that Machine Learning Algorithm makes.

These assumptions help the algorithm

1) to find the function that can map the inputs to the output,

2) to optimize the function in order to make good predictions.

For Example:

Linear Regression algorithm assumes linearity, normality, and homoscedasticity (equal variance)

Naive Bayes assumes that the data is Normally distributed, and conditional independence exists between the independent features.

K-NN makes the assumption that the data points close together will be similar, hence the new data point/unknown data will be placed with the majority of the neighbors. Support Vector Machines assume that the margins should be large

Thus, when an algorithm makes assumptions in order to model the data it can be understood as inductive bias.