Four colors, one remarkable proof
There is a simple-sounding question behind our name. How many colors do you need to color a map so that neighboring regions never share a color?
Four are enough. That is the four color theorem, a result from graph theory that took mathematicians more than a century to prove. Its statement is easy to understand. Its proof changed how we think about mathematics and computation.
The proof that needed a computer
In 1976, Kenneth Appel and Wolfgang Haken announced a proof built on mathematical reasoning and extensive computer checking. They had reduced an infinite family of possible maps to a finite collection of cases that could be examined computationally. The machines did not make the mathematics irrelevant; they made an otherwise extraordinary verification task possible.
It was a turning point. For one of the first times, a major mathematical theorem depended on computer work too extensive for a person to reproduce directly by hand. Other, more streamlined computer-assisted proofs followed, including a formally checked proof. A generally accepted proof that dispenses with computation altogether is still not known.
A simple question can demand more computation than any human could ever carry out.
From proving theorems to discovering algorithms
What links the theorem to 4colors Research is this: some computational problems are too intricate to navigate by intuition or manual design alone. And sometimes the methods we need to solve them are themselves difficult to discover.
Two algorithms can tackle the same problem and produce very different results within the same time. A better search strategy can reveal a feasible schedule, a more efficient allocation or a stronger circuit implementation that another method might never find in a practical time budget.
We work at that frontier. Mathematical optimization gives us structure. Machine learning can help guide search. Automated algorithm discovery and AI-assisted design help us explore promising new methods. Rigorous testing and benchmarking determine which ideas earn a place in our software.
The goal is not simply to make computation faster. It is to expand the range of problems that can be addressed well, and the decisions that can be improved as a result.
A name for the work ahead
The four color theorem shows how a simple question can hide a hard computation, and how computers can help answer it.
That is the idea behind 4colors Research: better algorithms, discovered and tested with care, opening more possibilities for the people who use them.
Explore our technologyMeet 4colors ResearchOur algorithms at work