Machine Learning vs. Numerical Computing

[_gnahz]

Hello,

I am an applied mathematics major who is intending on pursuing a MFE with the goal of ultimately becoming a quant.

Along with my applied mathematics major, I am planning on obtaining a CS minor. However, due to how popular the CS courses now are, we are limited to the number of senior-level courses we can take. This puts me in the position to have to decide between these courses. I have limited myself down to two areas that I believe will be most useful, Machine Learning and Numerical Computing. The course descriptions are as follows:

Machine Learning

1. Machine Learning and Data Mining - Fundamentals of supervised and unsupervised learning. Neural networks, policy gradient methods in reinforcement learning
2. Probabilistic Learning and Reasoning - Finding meaningful latent representations of data, taking advantage of large unlabeled datasets, and doing analogical reasoning automatically by building probabilistic models.

Numerical Computing

1. Numerical Algorithms - Learn how to develop accurate and efficient numerical algorithms for computing intensities for practical problems modelled by mathematics
2. Numerical Methods for Optimization Problems - Steepest descent, Newton's method, quasi-Newton methods, conjugate gradient methods and techniques for large problems

I believe that all four of these courses are taught with MATLAB. Any ideas/thoughts as to which pair of courses might be more useful in the future will be greatly appreciated. At this point in time, I am leaning towards the machine learning courses but I know that numerical computing/analysis is quite important for math students.

Thanks.

 

[Daniel Duffy]

NC, IMO

And then it is easy to move to ML.

Actually, a lot of ML is applied NC.

 

[bigbadwolf]

NC, hands down. ML is chic at the moment but NC is of proven worth and you'll also learn a fair bit about scientific coding.