3-D Print and Rubik's Cube-ify Almost Anything

As if a traditional Rubik's Cube wasn't hard enough, a new algorithm can turn any shape into a twisting puzzle and then create it on a 3D printer

2 min read
3-D Print and Rubik's Cube-ify Almost Anything
Image: Columbia University

A Rubik’s Cube is a 3-D puzzle designed to be enjoyed for 15 minutes, loathed 30 more minutes, and then placed in a drawer and forgotten. This is because the utility of a solved Rubik’s Cube is less than the utility of an unsolved Rubik’s Cube, so there is simply no motivation to solve it. 

But imagine if you could turn any object whatsoever into a puzzle that needs to be solved before you can use it. That would be fun, right? Sure it would, if by “fun” you mean “the worst.” So let’s do it!

Two computer science students from Columbia University have developed a method that allows people who have no idea what they're doing to create twisting 3D puzzles from arbitrary 3D models. Once you have the model in a computer, you can select your own rotation planes, and an algorithm will munch through everything, adjusting your model to prevent collisions and then 3D printing all the bits and pieces so that they interlock and rotate properly:

At the moment, this method works best with 3D models that have a large spherical component to their design, although the researchers are working on generalizing their technique. They’re also experimenting with ways of allowing some pieces to block the rotation of other pieces, which would allow for more rotational axes while also potentially enhancing the difficulty of the puzzle. From the sound of things, the biggest source of frustration has been the 3D printing itself, which makes assembly tricky and leaves the puzzles a bit fragile. So other potential improvements would be automatic generation of assembly instructions along with optimization of joint design for robustness.

Computational Design of Twisty Joints and Puzzles, by Timothy Sun and Changxi Zheng from Columbia University, will be presented at SIGGRAPH 2015 in Los Angeles, but you can read the full-length paper ahead of time here.

The Conversation (0)

The Future of Deep Learning Is Photonic

Computing with light could slash the energy needs of neural networks

10 min read

This computer rendering depicts the pattern on a photonic chip that the author and his colleagues have devised for performing neural-network calculations using light.

Alexander Sludds

Think of the many tasks to which computers are being applied that in the not-so-distant past required human intuition. Computers routinely identify objects in images, transcribe speech, translate between languages, diagnose medical conditions, play complex games, and drive cars.

The technique that has empowered these stunning developments is called deep learning, a term that refers to mathematical models known as artificial neural networks. Deep learning is a subfield of machine learning, a branch of computer science based on fitting complex models to data.

Keep Reading ↓ Show less