Optimization remains one of Google's biggest challenges. It helps teach machines how to learn from humans. This time around, the optimization techniques are not about organic search or paid search. It's about plasma fusion with atoms, not bits. Still, it all leads back to creating better algorithms for search and advertising.
Why? There are a bunch of smart people at Google trying to solve a bunch of complex problems that can lead to solutions to other complex problems.
Complex systems with many parameters are common in biology, physics, engineering, geology, and social science. Understanding and optimizing complex systems in biology, physics, engineering, geology and social science is difficult because of the time it takes with today's tools.
Google's researchers believe that progress can be made with new tools, which the company is trying to build. So to increase the speed of learning and optimization of plasma, Google's researchers developed the Optometrist Algorithm, which is motivated by human choices similar to search engines and ad serving.
"Our strategy of humans selecting between machine-generated settings is broadly applicable and differs from typical approaches to optimisation problems in the machine learning community, where strict figures of merit are automatically computed," per Google researchers. "The Optometrist Algorithm provides for efficient optimisation in areas where computation of such metrics is not possible."
The company says the algorithm combines the best of machine and human elements. The human element provides intuition in physics, while the machine searches high-dimensional space.
You can read more about the research and experiment here.