Previously, I had set up the framework / outline for the dissertation thesis. This weekend, I intend to work on Chapters 2 and 3. Chapter 1 covers the introduction and it is typically something to be tackled last. I need to get working on Chapter 2 and 3 over the next couple of weeks which includes the background and notation sections in the former, and the related work in the latter. This will help me completely narrow my focus down and develop the motivation section.
A big key component of Chapter 2/3 is to cover the general area of transfer learning and focus it down to my area of interest. This may be, at this time, a framework for boosting that works across the techniques of transfer; instance-based, parameter-based, etc.
A quick summary: Transfer Learning presupposes that an agent/algorithm’s performance on learning a task (or cost of) may be improved by transferring some knowledge from another task. Typically, this is a similar task, but a big idea is to discover how to map between tasks; i.e. learning a mapping between tasks and/or task domains. This, given some other work, would allow for transfer from different tasks in different domains. Some work in reinforcement learning and in learning transformations between images has been done in related research. This concept is important to real-world problems where training data is scarce or the cost of training data is high.