Research in Focus: Probablistic Programming

The traditional machine-learning approach relied on developing thousands of specific algorithms, each good for teaching machines how to do one or a few particular things. Chris Bishop, Distinguished Scientist, Microsoft Research Cambridge, is devising a new paradigm for solving machine learning, involving a model-based approach and probabilistic programming, which deals with the messy, complicated uncertainty of real-world data.

Watch Next