MIT Says Their New AI Programming Language Makes AI More Accessible to Everyone
As if we needed it to be easier for Skynet to be created. A team of MIT researchers has created an AI programming language that, they say, makes it easier for novices to start programming artificial intelligence.
Not only that, it will help experts further advance the field.
The new programming language is called 'Gen' and it is detailed in an MIT paper, titled Gen: a general-purpose probabilistic programming system with programmable inference.
In their paper, which was revealed at the Programming Language Design and Implementation conference last week, the researchers describe Gen, their new probabilistic-programming system.
Users can use Gen to write models and algorithms from multiple fields — such as computer vision, robotics, and statistics — where AI techniques are typically applied. They can do so without having to deal with deciphering complex equations or manually writing high-performance code.
Experts can also use the new programming language to write sophisticated models and inference algorithms that would have previously been too complex.
In the study, the researchers show that a short Gen program can infer 3-D body poses, a complex computer-vision inference task that can be applied in autonomous systems and augmented reality. While this is happening, the program can continue to fulfill other functions behind the scenes, an MIT statement explains.
Graphics rendering, deep-learning, and types of probability simulations are all processed while the user is utilizing the simpler programming language. This hugely reduces the time and skill needed to program AI.
AI for everyone?
Due to its relative simplicity, the MIT researchers say Gen can be easily used by anyone, whether they're novices or experts.
“One motivation of this work is to make automated AI more accessible to people with less expertise in computer science or math,” first author Marco Cusumano-Towner, a Ph.D. student in the Department of Electrical Engineering and Computer Science says in the MIT statement.
“We also want to increase productivity, which means making it easier for experts to rapidly iterate and prototype their AI systems.”
In a piece for the World Economic Forum, Eric Schmidt stated that AI should benefit the many, not the few. If AI programming becomes more accessible to everyone, that can only be a good thing — unless you believe it will lead to the rise of Skynet that is.