Many of us have, at some point, crumpled a piece of paper in frustration. But few of us have wondered; how does the way we crumple paper affect the ridges we impart on it? Besides paper, understanding how materials crumple can be relevant to a range of processes from the folding of biological membranes to the mechanical properties of thin films. 

We will attempt to answer some of these questions using two approaches. In the first, paper will be crumpled by hand, photographed and the ridges algorithmically extracted. Then, a convolutional neural network will be trained to predict the location of ridges from images of crumpled paper alone, circumventing the need for tedious experimentation. The second approach will involve using an existing code for simulating crumpling of bidimensional materials to investigate the effect of confining geometry on the resulting distribution of ridge lengths. In the final step, data from simulations will be combined with neural network supported experimental data to elucidate the difference between folding paper by hand (randomly) and folding paper in a controlled manner (simulation).

This project will require 2 interns, one focused on the experimental and neural network side, the other focused on simulations. The interns will communicate regularly to achieve the aims of this project.