A new course in computational biology at Carnegie Mellon University aims to give students a “deep understanding of the neural circuitry involved in chess and how they can be manipulated to solve games”.
The course, entitled “Brain Programming: Deep Neural Networks in Chess and Natural Language Processing”, was launched earlier this month by a team of students who include a professor from Carnegie Mellon’s Computer Science Department.
“This is the first time I’ve ever seen a course that is actually designed for a student who wants to learn how to program in the language of chess,” said Dr. David Ritchie, the course’s creator.
It’s a course which, he explained, is intended to teach the skills needed to “program the brain in the way you would expect a brain to do”.
“It is a way of teaching them to program their brains in the right way, and to do that, they need to understand the concepts of programming,” he said.
The students will also learn how “deep neural networks” (DNNs) are developed and programmed to understand chess moves.
DNN is a form of artificial intelligence, developed by a group of researchers at the University of Oxford in the United Kingdom.
A DNN is an artificial neural network that is capable of understanding and reasoning about the world around it.
Computer science professor and instructor of computer science Dr. Ritchie said that a DNN would be useful for “learning about the mind and understanding human nature”.
“If you have the right kinds of training and you understand the kinds of neural circuitry you need, then you can actually program your brain in a way that allows you to think about chess,” he added.
One of the main goals of the course is to “help students develop their abilities to use DNNs in the context of the brain”, said Dr Ritchie.
For the course, students will be asked to build a “fantastic chess AI” and use it to play games against other students.
“It’s not like you can just say, ‘Here’s a game of chess,'” said Ritchie about the course.
Instead, the students will need to write a program that would “explain how DNN works, how to build an AI using DNN and how to interact with the AI”.
“The aim is to help students understand how they’re doing this,” he explained.
In the course videos, students are introduced to neural networks (NNs), which are “very simple models of neural networks that have been trained to predict an object in the world”.
They are used to “generate new representations of an object”, which are then “reused in an algorithm” to produce new representations.
The first game of the semester will be against another student’s neural network, and the students are told that they are playing a game against a machine.
After a couple of moves, the AI “displays a game board, the game is over”.
“It was a really challenging course, but a lot of the students that came through really enjoyed it,” said Ritter.
During the course course, the participants will be given an “interactive training” which allows them to learn the programming language.
“The way it’s set up is that the first step is to build up an AI that can learn the language.
Then, after that, you can interact with it,” Ritchie explained.
In the end, Ritchie hopes that the course will give students “a deeper understanding of how the brain actually works”.
“What we’re really trying to do is get students to see how neural networks actually work,” he told Mashable.
“If you can learn how the mind works, then it’s a good place to start.”
For more on the course and how it’s being designed, head over to TechCrunch.