Machine Learning 10-701 2013/H2 Lecture 1

Introduction to Machine Learning (PhD level)

Basic overview of Machine Learning Techniques / Problems / Applications

[kapilgupta]

At 39:27 b^2 is added, what could be the good reason to add it, as the only w is involved in changing the function value, which if large for small changes in input features may not be robust of noise. But I could not get the good reason behind adding b^2 term