Active Predicate Learning
Planning in robotics environments is difficult in part due to continuous state and action spaces. One approach to this challenge is to use bilevel planning, where decisionmaking occurs in multiple levels of abstraction. However, the efficacy and efficiency of bilevel planning relies on the underlyin...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Published: |
Massachusetts Institute of Technology
2023
|
| Online Access: | https://hdl.handle.net/1721.1/150293 |
| _version_ | 1826193402550550528 |
|---|---|
| author | Li, Amber |
| author2 | Kaelbling, Leslie |
| author_facet | Kaelbling, Leslie Li, Amber |
| author_sort | Li, Amber |
| collection | MIT |
| description | Planning in robotics environments is difficult in part due to continuous state and action spaces. One approach to this challenge is to use bilevel planning, where decisionmaking occurs in multiple levels of abstraction. However, the efficacy and efficiency of bilevel planning relies on the underlying set of state and action abstractions. It is impractical to assume these abstractions as given (i.e. hand-designed by humans), so instead the agent should learn them, for example by exploring and interacting with its environment under the guidance of a teacher, from whom the robot may query expert knowledge. This is more difficult than the typical active learning problem setting because the robot must take actions to get to a state before it can make useful queries of the teacher about that state. This work develops an active learning framework for learning state abstractions (predicates) for effective and efficient task and motion planning. Given the names and arguments of the predicates in an environment and very few pre-labeled examples, the agent is able to learn predicate classifiers that enable it to successfully complete test tasks. |
| first_indexed | 2024-09-23T09:39:00Z |
| format | Thesis |
| id | mit-1721.1/150293 |
| institution | Massachusetts Institute of Technology |
| last_indexed | 2024-09-23T09:39:00Z |
| publishDate | 2023 |
| publisher | Massachusetts Institute of Technology |
| record_format | dspace |
| spelling | mit-1721.1/1502932023-04-01T03:53:10Z Active Predicate Learning Li, Amber Kaelbling, Leslie Silver, Tom Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Planning in robotics environments is difficult in part due to continuous state and action spaces. One approach to this challenge is to use bilevel planning, where decisionmaking occurs in multiple levels of abstraction. However, the efficacy and efficiency of bilevel planning relies on the underlying set of state and action abstractions. It is impractical to assume these abstractions as given (i.e. hand-designed by humans), so instead the agent should learn them, for example by exploring and interacting with its environment under the guidance of a teacher, from whom the robot may query expert knowledge. This is more difficult than the typical active learning problem setting because the robot must take actions to get to a state before it can make useful queries of the teacher about that state. This work develops an active learning framework for learning state abstractions (predicates) for effective and efficient task and motion planning. Given the names and arguments of the predicates in an environment and very few pre-labeled examples, the agent is able to learn predicate classifiers that enable it to successfully complete test tasks. M.Eng. 2023-03-31T14:45:39Z 2023-03-31T14:45:39Z 2023-02 2023-02-27T18:43:29.833Z Thesis https://hdl.handle.net/1721.1/150293 In Copyright - Educational Use Permitted Copyright MIT http://rightsstatements.org/page/InC-EDU/1.0/ application/pdf Massachusetts Institute of Technology |
| spellingShingle | Li, Amber Active Predicate Learning |
| title | Active Predicate Learning |
| title_full | Active Predicate Learning |
| title_fullStr | Active Predicate Learning |
| title_full_unstemmed | Active Predicate Learning |
| title_short | Active Predicate Learning |
| title_sort | active predicate learning |
| url | https://hdl.handle.net/1721.1/150293 |
| work_keys_str_mv | AT liamber activepredicatelearning |