Taking the magic out of machine learning
Artificial intelligence (AI) and its subcomponent machine learning (ML) are almost everywhere in industrial processes, and cement manufacturing is no exception.1,2 Despite the common use of these terms, how a model is built and implemented remains a mystery to many. Holcim Innovation Center demystifies the process by stepping through a machine learning project. By Louise Harding and Xavier Cieren, Holcim Innovation Center, France.
Figure 1: machine learning implementation steps
For newcomers to the domain, machine learning (ML) can appear to be an opaque process. However, the preparation of an ML model is not as hands-off as passing your data through special algorithms on highly-powered computers and fine-tuning the results. It is a coordinated, sometimes challenging, task that obtains the best results when domain experts and data scientists work together.2
