Machine Learning Through Probabilistic Programming

摘要

Programming that's probabilistic? Really? That doesn't make much sense... Or that's what I thought when I started to work in this domain. The researchers I was listening to didn't have the traditional view of placing machine learning (ML) problems into categories. Instead, they just expressed the real-world processes that lead to the generation of their data in a computer-readable form. That's what they call a model and its representation-a probabilistic program.