The Turing Test for Graph Drawing Algorithms

TL;DR

This paper investigates whether graph drawing algorithms can produce outputs indistinguishable from human-drawn graphs through a human-centered experiment, revealing that some algorithms, especially force-directed and MDS, can pass as human drawings.

Contribution

It introduces a Turing Test framework for graph drawing algorithms and evaluates their ability to mimic human drawing quality on small graphs.

Findings

Force-directed and MDS algorithms often pass as human-drawn graphs.

Hand-drawn graphs are generally judged higher quality than algorithm-generated ones.

Algorithm performance varies with graph size and type.

Abstract

Do algorithms for drawing graphs pass the Turing Test? That is, are their outputs indistinguishable from graphs drawn by humans? We address this question through a human-centred experiment, focusing on `small' graphs, of a size for which it would be reasonable for someone to choose to draw the graph manually. Overall, we find that hand-drawn layouts can be distinguished from those generated by graph drawing algorithms, although this is not always the case for graphs drawn by force-directed or multi-dimensional scaling algorithms, making these good candidates for Turing Test success. We show that, in general, hand-drawn graphs are judged to be of higher quality than automatically generated ones, although this result varies with graph size and algorithm.