We at the Human-Centered AI Lab are working on our FWF-project-P-32554 to systematically develop a ground truth data set for visual concept learning. The goal of this experiment is to capture human descriptions of patterns with all possible human explanations one can imagine. We simply want to get some insight into the various types of descriptions/explanations that different people come up with – it is about the diversity of explanatory patterns.
Disclaimer: This experimental survey is voluntary, anonymous, for scientific investigation only, and raises no ethical issues.
Please take the following very short survey – it will take you less than 10 minutes, just fill in as many as possible explanations into the boxes:
Here you find some background reading:
https://arxiv.org/abs/2103.00519
