Statistical Machine Learning (SML) – Graduate Degree Specialization

Why take the SML program?

If you are a computer science student, interested in machine learning, why should you take the SML program? What are the benefits? And what are the pitfalls?

Fact is, 90 percent of machine learning is based on statistics ideas. Statistical ideas, statistical thinking constitute the core of the subject. If you really want to understand overfitting, cross validation and what is it good for, what are the limits to learnability, what are the adaptive methods, why is LASSO a good idea (if it is a good idea at all), you better speak the language that is used to describe results about these issues. The SML program gives you a chance to build strong foundations in probability theory and statistics. These days the boundary between machine learning and statistics is even less clear than it was ever before. Statisticians publish in machine learning journals and at machine learning conferences and machine learning researchers publish at statistics journals. After all, they are all after creating better techniques to deal with the need to create better models to produce better predictions. In fact, the demand for rigorous analysis of algorithms is bigger than ever -- and for all the good reasons: A solid understanding of algorithms is necessary so that the tower of results building on the top of each other do not collapse. Empirical evidence is important, but it can never tell the whole story. The reasoning so far tried to convince you that taking the program enhances your chances of building a solid foundation. This might even lead to a good PhD. But what happens after that? You will surely look for jobs. Will you be at an advantage or disadvantage compared to others who do not have a degree specialization? We think that taking the program puts you at an advantage. In fact, it almost doubles your chances of being employed. These days, even employers (looking for machine learning researchers) are aware that machine learning and statistics are tightly intervowen. They will likely care about that you have a degree that certifies that you speak both languages. You can also think that now you can apply for both jobs that require a computer science/machine learning background or a statistics/probability theory background. You can double your chances. If you consider staying in the academia, you should know that many, many openings are in statistics. In fact, quite a few graduates of us got jobs at statistics departments. Your chances would be even better if you took the SML program.

What if you are a statistics student? Why should you care?

Well, machine learning is a very vibrant, fast growing part (if you like) of statistics. Many new models pop up, the opportunities flourish. For some reason, in machine learning people like non-standard models, situations. This creates many wonderful research opportunities. Also, being a new subject, maybe you do not even have to work that hard to get a reasonable recognition by the community (well, the truth is that everyone has to work hard).

Regarding the job situation, the above reasoning works for you, too. You can double your chances. Having a specialization in machine learning means that you are literal in computer science. Google, Yahoo, Amazon, Netflix and more: Probably you have heard about these companies. They take many of the computer science graduates and they are especially interested in employing machine learning researchers.

Who should not take the SML program?

If you are a computer science student and you are bored of theories, math, probability theory, do not take the program. If you are afraid of hard work, do not take it. In fact, the load in this program is slightly higher than in the normal program. But this should be no surprise: there are no free lunches in graduate school (well, free pizza, maybe..). For more specific information about the program (requirements, etc.), visit the program description page here.