[URL moved!] Contribute on photo rating data set

[Xiao, Han]

[Please kindly ignore it if you have done the rating work]

Hi all,

We thank those people who contributed on this work.

The photo rating system is now moved to a new URL. Several issues and bugs reported
by users have been resolved. Please visit http://ml.sec.in.tum.de/PhotoStayer/ and give it a try.

The goal is to collect ratings of 100 photos from 100 different people, which we

call it TUM100² data set. The data set will be used to train our recently

proposed crowdsourcing model, which can learn the ground truth from multiple

responses of unreliable experts (crowds). In this experiment, we expect our

model can learn to determine the artistic charm of a photo.

To contribute on this data set,

1. Please first click ``get an ID'' for a simple registration.

2. There will be a short introduction that tells you some background about this

system and the basic concept of the experiment.

3. We'd appreciate it if you can rate at least 100 photos.

4. Our model is developed to be resilient to the adversarial noise. Thus, if you

wish you can be a hater who dislike every photo, or casually give random

ratings. Please give a low confidence value in the survey when you do so.

5. The data set will be open to the public in the future, yet your personal information will
never be distributed.

Thanks in advance,

Sincerely,

Han

--
Han Xiao, Ph.D. Candidate
Technische Universität München, Informatics, I20
Boltzmannstr 3, 85748 Garching
Tel: +49.89.289.18590
h.x...@tum.de
http://home.in.tum.de/~xiaoh