Back-to-Basics Weekend Reading - Machine Learning
Machine learning is a scientific discipline that explores the construction and study of algorithms that can learn from data. Such algorithms operate by building a model from example inputs and using that to make predictions or decisions, rather than following strictly static program instructions.
Machine Learning is playing an increasing important role in many areas of our businesses and our lives. ML is used for predictive analytics and predictive modeling, e.g. making predictions about the likelihood that a certain event is going to happen (will this customer be interested in this item, is this message spam). At Amazon machine learning has been key to many of our business processes, from recommendations to fraud detection, from inventory levels to book classification to abusive review detection. And there are many more application areas; search, autonomous cars (and drones), text and speech recognitions, game play, etc.
As is the case with most computer science, Machine Learning is not new. It roots are in the late 50’s early 60’s, although of course one can even claim that Turing was the first to discuss the topic. For this weekends reading instead going back to the early days I have picked two survey papers on two major categories of machine learning: supervised and unsupervised learning.
But first I suggest you read professor Pedro Domingos paper to understand the context of machine learning and what the prerequisites are for it to be successful.
A Few Useful Things to Know about Machine Learning, Pedro Domingos, Communications of the ACM, 55 (10), 78-87, 2012.
Unsupervised machine learning:
Data clustering: a review, A.K. Jain, M.N. Murty, and P.J. Flynn, ACM Computer Surveys, 31, 3 (September 1999)
Supervised machine learning:
Machine learning: a review of classification and combining techniques, S. B. Kotsiantis, I. D. Zaharakis, and P. E. Pintelas, Artificial Intelligence Review 26:159–190 (2006)
If all of this gets you excited and want to learn more I suggest you take professor Domnigos class on Machine learning at coursera:
Machine Learning - Why write programs when the computer can instead learn them from data? In this class you will learn how to make this happen, from the simplest machine learning algorithms to quite sophisticated ones. Enjoy!
For those of you who are interested in a more popular treatment of prediction I suggest you read Nate Silver’s book The Signal and the Noise: Why So Many Predictions Fail–but Some Don’t