Personalized Estimation of Engagement From Videos Using Active Learning With Deep Reinforcement Learning.
Ognjen (Oggi) RudovicHae Won ParkJohn BuscheBjörn W. SchullerCynthia BreazealRosalind W. PicardPublished in: CVPR Workshops (2019)
Keyphrases
- active learning
- reinforcement learning
- learning algorithm
- learning process
- crowd sourced
- supervised learning
- transfer learning
- semi supervised
- learning environment
- relevance feedback
- batch mode
- estimation algorithm
- function approximation
- pool based active learning
- active exploration
- e learning
- online video
- experimental design
- learning problems
- machine learning
- learning strategies
- video sequences
- human activities
- video frames
- model free
- parameter estimation
- multi agent
- selective sampling
- user engagement
- dynamic programming
- active learning strategies
- training set