Personalized Estimation of Engagement From Videos Using Active Learning With Deep Reinforcement Learning
© 2019 IEEE. Perceiving users' engagement accurately is important for technologies that need to respond to learners in a natural and intelligent way. In this paper, we address the problem of automated estimation of engagement from videos of child-robot interactions recorded in unconstrained env...
| Main Authors: | , , , , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
Institute of Electrical and Electronics Engineers (IEEE)
2021
|
| Online Access: | https://hdl.handle.net/1721.1/137137 |
| _version_ | 1826215998798168064 |
|---|---|
| author | Rudovic, Ognjen Park, Hae Won Busche, John Schuller, Bjorn Breazeal, Cynthia Picard, Rosalind W. |
| author2 | Massachusetts Institute of Technology. Media Laboratory |
| author_facet | Massachusetts Institute of Technology. Media Laboratory Rudovic, Ognjen Park, Hae Won Busche, John Schuller, Bjorn Breazeal, Cynthia Picard, Rosalind W. |
| author_sort | Rudovic, Ognjen |
| collection | MIT |
| description | © 2019 IEEE. Perceiving users' engagement accurately is important for technologies that need to respond to learners in a natural and intelligent way. In this paper, we address the problem of automated estimation of engagement from videos of child-robot interactions recorded in unconstrained environments (kindergartens). This is challenging due to diverse and person-specific styles of engagement expressions through facial and body gestures, as well as because of illumination changes, partial occlusion, and a changing background in the classroom as each child is active. To tackle these difficult challenges, we propose a novel deep reinforcement learning architecture for active learning and estimation of engagement from video data. The key to our approach is the learning of a personalized policy that enables the model to decide whether to estimate the child's engagement level (low, medium, high) or, when uncertain, to query a human for a video label. Queried videos are labeled by a human expert in an offline manner, and used to personalize the policy and engagement classifier to a target child over time. We show on a database of 43 children involved in robot-assisted learning activities (8 sessions over 3 months), that this combined human-AI approach can easily adapt its interpretations of engagement to the target child using only a handful of labeled videos, while being robust to the many complex influences on the data. The results show large improvements over a non-personalized approach and over traditional active learning methods. |
| first_indexed | 2024-09-23T16:40:46Z |
| format | Article |
| id | mit-1721.1/137137 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T16:40:46Z |
| publishDate | 2021 |
| publisher | Institute of Electrical and Electronics Engineers (IEEE) |
| record_format | dspace |
| spelling | mit-1721.1/1371372024-08-09T19:44:24Z Personalized Estimation of Engagement From Videos Using Active Learning With Deep Reinforcement Learning Rudovic, Ognjen Park, Hae Won Busche, John Schuller, Bjorn Breazeal, Cynthia Picard, Rosalind W. Massachusetts Institute of Technology. Media Laboratory © 2019 IEEE. Perceiving users' engagement accurately is important for technologies that need to respond to learners in a natural and intelligent way. In this paper, we address the problem of automated estimation of engagement from videos of child-robot interactions recorded in unconstrained environments (kindergartens). This is challenging due to diverse and person-specific styles of engagement expressions through facial and body gestures, as well as because of illumination changes, partial occlusion, and a changing background in the classroom as each child is active. To tackle these difficult challenges, we propose a novel deep reinforcement learning architecture for active learning and estimation of engagement from video data. The key to our approach is the learning of a personalized policy that enables the model to decide whether to estimate the child's engagement level (low, medium, high) or, when uncertain, to query a human for a video label. Queried videos are labeled by a human expert in an offline manner, and used to personalize the policy and engagement classifier to a target child over time. We show on a database of 43 children involved in robot-assisted learning activities (8 sessions over 3 months), that this combined human-AI approach can easily adapt its interpretations of engagement to the target child using only a handful of labeled videos, while being robust to the many complex influences on the data. The results show large improvements over a non-personalized approach and over traditional active learning methods. 2021-11-02T17:35:59Z 2021-11-02T17:35:59Z 2019 2021-06-24T15:15:56Z Article http://purl.org/eprint/type/JournalArticle https://hdl.handle.net/1721.1/137137 Rudovic, Ognjen, Park, Hae Won, Busche, John, Schuller, Bjorn, Breazeal, Cynthia et al. 2019. "Personalized Estimation of Engagement From Videos Using Active Learning With Deep Reinforcement Learning." IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, 2019-June. en 10.1109/CVPRW.2019.00031 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf Institute of Electrical and Electronics Engineers (IEEE) MIT web domain |
| spellingShingle | Rudovic, Ognjen Park, Hae Won Busche, John Schuller, Bjorn Breazeal, Cynthia Picard, Rosalind W. Personalized Estimation of Engagement From Videos Using Active Learning With Deep Reinforcement Learning |
| title | Personalized Estimation of Engagement From Videos Using Active Learning With Deep Reinforcement Learning |
| title_full | Personalized Estimation of Engagement From Videos Using Active Learning With Deep Reinforcement Learning |
| title_fullStr | Personalized Estimation of Engagement From Videos Using Active Learning With Deep Reinforcement Learning |
| title_full_unstemmed | Personalized Estimation of Engagement From Videos Using Active Learning With Deep Reinforcement Learning |
| title_short | Personalized Estimation of Engagement From Videos Using Active Learning With Deep Reinforcement Learning |
| title_sort | personalized estimation of engagement from videos using active learning with deep reinforcement learning |
| url | https://hdl.handle.net/1721.1/137137 |
| work_keys_str_mv | AT rudovicognjen personalizedestimationofengagementfromvideosusingactivelearningwithdeepreinforcementlearning AT parkhaewon personalizedestimationofengagementfromvideosusingactivelearningwithdeepreinforcementlearning AT buschejohn personalizedestimationofengagementfromvideosusingactivelearningwithdeepreinforcementlearning AT schullerbjorn personalizedestimationofengagementfromvideosusingactivelearningwithdeepreinforcementlearning AT breazealcynthia personalizedestimationofengagementfromvideosusingactivelearningwithdeepreinforcementlearning AT picardrosalindw personalizedestimationofengagementfromvideosusingactivelearningwithdeepreinforcementlearning |