Machine learning (ML) and artificial intelligence (AI) were once concepts relegated to only the most optimistic observers, much like self-driving electric vehicles and smartphones once were. But if it isn't obvious, the times have changed. Today, ML and AI--along with the immensely powerful data collection and analytics tools that power those processes--are a mainstay of modern life. Every day, people interact with products and services powered by some of the world's most ground-breaking technology. In the financial sector specifically, ML and AI present an enormous opportunity to institutions to revolutionise their businesses and generate both top- and bottom-line results.

We propose a technique for decomposing and compressing the dataset in the parameter learning problem in Markov random fields. Our technique applies to incomplete datasets and exploits variables that are always observed in the given dataset. We show that our technique allows exact computation of the gradient and the likelihood, and can lead to orders-of-magnitude savings in learning time.