This new workpackage brings together the activities covered by the old workpages WP8  Machine Learning for Multimedia Content  and WP7 Computation.

The objectives of this workpackage include:
(1) development of new ML techniques specific for multimedia understanding;
(2) extension and integration of existing ML techniques; and
(3) coordination of the application of ML methods by the network to different tasks of multimedia understanding.

The workpackage is broken down into 6 tasks: 

 

Task 1.  Supervised Learning
Task 2.  Kernel Methods
Task 3.  Dimension Reduction and Feature Transformation
Task 4.  Semi-Supervised Learning & Active Learning
Task 5.  Applications of ML in Multimedia Processing
Task 6.  Book Project.

Tasks 1-5 represent 5 important subareas within machine learning that are relevant to processing multimedia content.
Tasks 2 is a new task in this workpackage reflecting the importance of Kernel Methods in the application of Machine Learning techniques in processing Multimedia Content and in the interest in these techniques amongst Muscle Partners.
Task 6 reflects work that is already underway but is not more formally organized in this task.

 

In addition, there are 3 e-teams associated with this workpackage, where most of the work will take place in JPA 4 (the last year of the project).  Please see the separate e-team page for more details.