Privileged Machine Learning for Prediction

Machine learning for prediction suffers from asymmetric distribution, such as posterior information, future information and hidden information. With some additional information only available in training, how to learn a machine learning model with them remains a key challenge. Despite recent advances in important domains such as vision and medicine, the standard learning under privileged information paradigm does not offer a satisfactory solution for learning variational privileged information. This thesis introduces how to learn under variational privileged information by leveraging asymmetric distribution and machine learning algorithms, specifically via 1) semi-supervised learning, 2) adversarial learning, 3) multi-task learning, 4) slack variable learning in support vector regression.

Author dc:creator

• Shu, Yangyang

Statement dc:rights

• The author owns the copyright in this thesis including all reproduction and reuse rights for the work. The work may not be altered without the permission of the copyright owner. Attribution is essential when quoting or paraphrasing from this thesis.
• au.edu.uts.lib/ppc
• info:eu-repo/semantics/openAccess

Language dc:language.iso

en_US

Handle dc:identifier.uri

http://hdl.handle.net/10453/154754

OAI identifier oai:identifier

oai:opus.lib.uts.edu.au:10453/154754