Coursera Structuring Machine Learning Projects Certification

Having completed the third course of's Deep learning specialization: Structuring Machine Learning Projects Certification, I was able to gain insight on many important concepts to which not much attention is often given, but are nonetheless crucial in implementing effective machine learning systems, such as orthogonalization, metrics selection appropriate to the problem, satisficing vs. optimizing metrics, the correct split of training/dev/test sets, analysis of error including avoidable bias, training/dev/set errors, the notion of Bayesian optimal error and "super-human" performance, ways to analyze and deal with bias, variance, and data mismatch problems, and important paradigms such as multi-task learning, transfer learning, end-to-end learning. 


#machinelearning #deeplearning #transferlearning #orthogonalization #endtoendlearning #multitasklearning #structureingmlprojects


Popular posts