Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

This seems like a nice and useful method, but the conclusions of the paper seem problematic, specifically that the resulting visualizations are more "effective" and higher "quality" than t-SNE (or anything else for that matter). Can that aspect of a visualization method be judged by only showing a few examples, without any tests of how people actually use or benefit from it?


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: