S3 Home‎ > ‎

Our Work

Computers and computation are different things. Unfortunately, the similarity of the words causes confusion. Computation uses computers to apply mathematics to problem solving. Computers are in there, but it's the math that solves the problems. Knowing how to configure networks and administer machines is important, and we do it for our own networks and machines, but that has little to do with computation, which is the thing we do to give our customers value. At S3, each of us has spent the majority of our lives figuring out how to solve real problems using engineering, science, and math, and we are comfortable making computers do as much of that work as possible. 

Because of our unique experience, we bring data science and simulation science together. And we think this matters a lot. Data science, and particularly "Big Data", develop mathematical models that correlate well to the real world but usually don't address causation. Usually it's important to know why the world is like that. Simulation science produces causal models that may or may not correlate well with the real world. In most problems, both factors are important. As an example, data science methods might produce a classifier to decide whether a college applicant is likely to be a successful student. That classifier might correlate very well to past student performance. And that classifier might include totally arbitrary factors such as whether the loan applicant drives a blue car versus a red car. And if we apply that classifier to the selection process, it can adversely affect applicants and change their behavior in costly ways. They may trade in their red cars for blue cars to increase their likelihood of acceptance. This would be a bad thing, and could further entrench historical prejudices. In a surprising number of cases, we end up building simulations to prevent reliance on spurious correlation. Sometimes the data science comes first, sometimes the simulation science comes first, sometimes they both proceed together. But the objective is always to produce models of the world that make sense.
One word cloud for the thesis or dissertation of each S3 principle investigator.