Predictive models from our researchers, at your service
In the age of artificial intelligence and deep learning, making predictive models can make or break a business. And our IBM Research team specialise in finding AI solutions to real-world industrial problems.
Bayesian optimisation is a way to create those predictive models automatically. It's used to solve a variety of problems in fields as varied as robotics, animation, planning, and experimental particle physics.
Our researchers have created Bayesian Optimisation as a Service. We gave it an acronym, obviously: BOaaS.
We've demonstrated it working very efficiently in high performance computing and experimental design, as well as reducing the potential for bias and inefficiency.
You could even say that our algorithm understands when it can trust itself, and when it needs to seek more information. Think of it as "AI for AI."
BOaaS effectively takes the human out of the loop, so that you don't need people to become experts in a complex mathematical discipline.
It's a way for people to use state-of-the-art automated model tuning, all from a very simple web user interface. Calculate like a BOaaS!