nickboucher/imperceptible

Do you have the code/model you used to generate figure 4?

Closed this issue · 2 comments

I am trying to visually identify homoglyphs and I was wondering about the details of models/hyperparameters you use to create figure 4. I needed stronger homoglyph mappings than the official unicode one... You mentioned that well-tuned unsupervised clustering algorithms can help the process, can you provide some guidance on this? Thanks!

Found it, nvm. Thank you for the great work

Glad to hear! If there's anything else you have trouble locating later, please feel free to re-open this issue.