Some like it hot - visual guidance for preference prediction · HackerPapers discussions

Some like it hot - visual guidance for preference prediction — mentioned by timofter

Wed, 06 Jan 2016 13:00:16 GMT

Hi fatjokes, I am Radu, one of the authors. As mentioned in our article (http://arxiv.org/abs/1510.07867v1), the training dataset for attractiveness is from Blinq. The underage and above 37 years old face images were discarded. All people are heterosexual in our training dataset and mainly from Zurich area, Switzerland. Therefore, our model of attractiveness fits the cultural bias of Zurich, Switzerland. We consider faces.ethz.ch a little fun tool. I hope that the fans of Yoona will understand :) With more training data from abroad Switzerland our algorithms will fit their expectations. For…

Paper and discussion history. A mention is not an endorsement.

Some like it hot - visual guidance for preference prediction — mentioned by MrQuincle

Tue, 05 Jan 2016 23:17:20 GMT

Of course I tried a bunch of pictures of myself, just as everybody at HN is doing, ha ha! I didn't get above the nice level. First picture of my wife she immediately got the godlike level. Unfair advantage!! "Before performing any experiments we removed underage people, anyone over 37 and bi- and homosexual users as these comprise only a small minority of the dataset." That's already biasing it of course! "Interestingly, 44.81% of the male ratings are positive, while only 8.58% of the female ratings are positive. Due to this strong bias of ratings by women, we only predict the ratings of…

Paper and discussion history. A mention is not an endorsement.

Some like it hot - visual guidance for preference prediction — mentioned by MrQuincle

Tue, 05 Jan 2016 22:17:45 GMT

Hi Rasmus, you described your paper on age estimation on the site. Do you also have something written on the hotness scale? :-) Edit: Sorry, missed it! http://arxiv.org/pdf/1510.07867v1.pdf

Paper and discussion history. A mention is not an endorsement.