Co-speech Gesture Synthesis by Reinforcement Learning with Contrastive Pretrained Rewards.
Mingyang SunMengchen ZhaoYaqing HouMinglei LiHuang XuSongcen XuJianye HaoPublished in: CVPR (2023)
Keyphrases
- reinforcement learning
- multimodal interfaces
- hand movements
- speech recognition
- markov decision processes
- function approximation
- multi stream
- reinforcement learning algorithms
- hidden markov models
- gesture recognition
- human computer interaction
- reward function
- optimal policy
- state space
- sign language
- automatic speech recognition
- hand gestures
- audio visual
- facial animation
- speech synthesis
- speech signal
- function approximators
- neural network
- multiarmed bandit
- reward shaping
- model free
- machine learning
- multi agent
- texture synthesis
- action selection
- markov decision problems
- total reward
- text to speech
- learning mechanism
- multimodal interaction
- noisy environments
- spoken language
- facial expressions
- mobile robot
- user interface
- continuous stream
- control policy
- policy iteration