Machine Learning System to Learn what Users Find Attractive in On-line Photographs

In this presentation I will present a machine learning system that learns what computer users find attractive in profile images (containing frontal face shots) from the web. Once the system is trained to a given user it will then be able to predict that user's future attractiveness rating (on a scale of 1-10) of profile images from the web. This technology has numerous applications, one example being On-line dating websites that would be able to match couples that would find each other's images attractive in addition to just profile matching.