The Defense Talk
August 17, 2012
Start this low-fi video and then follow along using the slides below.
The Dissertation Document
W. Bradley Knox. Learning from Human-Generated Reward. Dissertation. September 2012. [pdf]
Learning from Human-Generated Reward
The Research Question
How and to what extent can agents harness the information contained in human-generated signals of reward to learn sequential decision-making tasks?
Defense slides:
For a visual map of the contents of this dissertation and how each chapter relates to current and future publications, refer to slide 6 above from the defense talk. Moving from the leftmost box in an inverted ‘V’ sequence, the boxes correspond to Chapters 2, 3, 4, 5, and then 6.