As I look back on the last ten years of technological progress, I am stricken by the
number of breakthrough technologies that lie in the intersection of mathematics and
computer science. Of course, this isn't to say that other computer science achievements
were orthogonal to mathematics--after all, computer science is, in reality, a field of
applied mathematics--only that the developments that combine purely mathematical fields
like graph theory and statistics with the unrivaled computing power of the 21st century
are unique.
In the same way that living matter and chemistry allow humans to make decisions, be
creative, learn, and understand, silicon and electricity empower the computer. We may
not understand how to recreate our own intelligence, but we certainly understand the
binary simplicity of our semi-intelligent creations. But what role will mathematics play
in the future of computing? My guess is that it will be larger than ever.
Five years ago, the only companies invested in hiring mathematically-inclined computer
scientists and mathematicians were Google, Amazon, the NSA, and a select few other
large, well-funded companies. Only a few years ago, starting a company that made heavy
use of applied mathematics was impractical. Today, the landscape has changed. Open
source machine learning libraries have made basic machine learning applications
available to the hobbyist, the college start-up founder, and the aspiring engineer.
Where will this trend head? While I cannot say for sure, my guess is that the prominence
of mathematicians and mathematics in computer science will only grow. Machine learning
and neural networks are only the beginning. These fields will undoubtedly grow, with
more specific subfields being developed over time, as is already happening.
But there's a problem. The mathematics currently used for computer science AI
applications is inelegant. While we are currently satisfied with simplistic Bayesian
models applied to massive quantities of data, that will not remain the case much longer.
In the past, quality data in high quantities was hard to come by. Today, data is no
longer the limiting factor; but soon, mathematics may be.
The best thing a university-bound or enrolled computer science student can do is develop
a strong background in mathematics. Computer science with a heavy dose of mathematics is
the next ten years of computing. We may not be able to make human-intelligent computers,
but we can certainly create Martian-intelligent computers.