Game uses human intuition to assist computers in solving knotty problems

If there is one core issue underlying hardware design tasks, it’s that the computers assigned to carry them out do not have human intuition and visual pattern recognition skills for creating optimal designs. Enter FunSAT, an online logic puzzle that lets integrated circuit (IC) designers arrange and select transistors and their connections on silicon microchips.

We have reached a stage of advancement where humans need not spend time over mundane tasks such as chip architecture. They can instead focus on the larger picture involving end users, be they people or machines. Computers are fully capable of designing small and powerful microchips to meet our technological needs.

But Artificial Intelligence can only flip through possible arrangements while determining where transistors and their connections may be placed on an IC. Computers do not have an instinct for recognizing optimal design or even what constitutes a better one. Here’s where FunSAT, a human-computing inspired game to solve Boolean satisfiability problems, comes in.

This prototype has been developed by University of Michigan computer science researchers Andrew DeOrio and Valeria Bertacco. And you don’t even have to be a computer scientist to play FunSAT! Liking puzzle games such as Sudoku is more than enough.

As Bertacco explains, “Computer games can be more than a fun diversion. Humans are good at playing games and they enjoy dedicating time to it. We hope that we can use their strengths to improve chip designs, databases and even robotics.”

The FunSAT board features columns and rows of red, gray and green bubbles in various sizes. Buttons present around the perimeter can be turned blue or yellow. The color of a button determines the color of bubbles on the board. The game’s objective is to use the buttons present on the perimeter to toggle all the bubbles green.

You can contribute to the design of complex computer systems by solving difficult problems on the FunSAT board. Once solved, a computer scientist can simply look at the color of each button to figure out the answer to a specific problem. The game features some classic and highly complicated mathematical questions that involve picking the most optimal arrangement.

The player must assign a set of variables to the right true or false categories to fulfill all the basic constraints of the problem. These constraints are represented by bubbles while the variables depict by the perimeter buttons. Bubbles become green when they are satisfied and perimeter buttons are assigned to true or false when players click the mouse to make them yellow (true) or blue (false).

Clues about what to do next are given when the player right-clicks a bubble. A large bubble controls more buttons and each of these buttons affect other bubbles simultaneously and in different ways. A button that converts several bubbles to green is also bound to turn other bubbles from green to gray, or even red.

The game prototype has been implemented in Java by University of Michigan undergraduate, Erica Christensen, and is free to play on FunSAT’s official website.