What is Mapped onto a Latent Variable
Type: Bachelor CS/EE
If you are interested please contact :
In this assignment, we want to analyse what type of variations of the input of a neural network lead to the same output. This is important, because sometimes this behaviour is desired, but at other times, it is not because it is unpredictable. An example of a desired mapping of input variations to the same output is robustness to variations of pose or illumination of a facial image in face recognition. This is a desired behaviour, because we don’t want the pose or the illumination to affect the recognition result.