Other scripts · 2022
Letting the computer decide which shapes are the same sign
Corazza and colleagues · PLOS ONE (Cypro-Minoan)

In a sentence
Images plus context sorted Cypro-Minoan’s variant shapes into signs.
In plain words
Scholars disagree about whether two similar shapes are one sign or two. For Cypro-Minoan, this team let a model look at each sign’s picture and where it appears, and group them. The Indus script has the same problem: the two main sign lists count about 400 and about 700 signs.
Technical detail
Sign2Vec: unsupervised clustering on image features plus distributional context, validated on Cypro-Minoan with known-answer checks.
What we take from it
Testing nowTesting now. We first try the method on scripts whose true sign list we know. If it cannot rebuild those lists, we will not trust what it says about the Indus list.
The idea we are borrowing
Build the sign list from pictures plus context, to settle about 400 vs about 700 signs
Where it stands: Testing now.