The new probabilistic programming languages
Probabilistic programming can do in 50 lines of code what used to take thousands of lines of traditional programming
Computer scientists are beginning to develop systems based on so-called "probabilistic programming languages."
The probabilistic approach to programming is like straying away from mathematical thinking, and moving onto a more intuitive approach. This will make machine-learning applications easier to build, according to the researchers.
Next June, at the Computer Vision and Pattern Recognition conference, MIT researchers will demonstrate that, on some standard computer-vision tasks, systems written in a probabilistic programming language are competitive with conventional systems, but consist only in a small fraction of the lines of code. Probabilistic programming can do in 50 lines of code what used to take thousands of lines of "traditional" programming.
"This is the first time that we're introducing probabilistic programming in the vision area," says Tejas Kulkarni, an MIT graduate student in brain and cognitive sciences. "The whole hope is to write very flexible models, both generative and discriminative models, as short probabilistic code, and then not do anything else. General-purpose inference schemes solve the problems."
"When you think about probabilistic programs, you think very intuitively when you're modeling," Kulkarni says. "You don't think mathematically. It's a very different style of modeling."