Off-campus UMass Amherst users: To download campus access dissertations, please use the following link to log into our proxy server with your UMass Amherst user name and password.

Non-UMass Amherst users: Please talk to your librarian about requesting this dissertation through interlibrary loan.

Dissertations that have an embargo placed on them will not be available to anyone until the embargo expires.

Author ORCID Identifier

https://orcid.org/0000-0001-8360-9214

Document Type

Open Access Dissertation

Degree Name

Doctor of Philosophy (PhD)

Degree Program

Computer Science

Year Degree Awarded

2019

Month Degree Awarded

September

First Advisor

Roderic A. Grupen

Second Advisor

Joydeep Biswas

Third Advisor

Philip S. Thomas

Fourth Advisor

Weibo Gong

Subject Categories

Artificial Intelligence and Robotics | Computer Sciences

Abstract

Deep learning has been successful in a variety of applications, such as object recognition, video games, and machine translation. Deep neural networks can automatically learn important features given large training datasets. However, the success of deep learning in robotic systems in the real world is still limited mainly because obtaining large datasets and labeling are costly. As a result, much of the successful work in deep learning has been limited to domains where large datasets are readily available or easily collected. To address this issue, I propose a framework for acquiring re-usable skills efficiently combining intrinsic motivation and the control basis framework --- a developmental architecture implemented using a landscape of attractors. A deep neural classifier is used to predict probabilistic control affordances representing controller states accessible by way of actions. Information theoretic motivation with embedding distance influences exploration in a way that supports learning control affordances efficiently. The learned affordance is used as a filter to help a robot experience more positive outcomes of controllers when a robot learns a new affordance. I conduct quantitative experiments using a dynamic simulator and the results show that the proposed learning framework enables a mobile manipulator to learn deep sensorimotor skills more efficiently in a self-supervised learning manner.

Creative Commons License

Creative Commons Attribution 4.0 License
This work is licensed under a Creative Commons Attribution 4.0 License.

Share

COinS