SAC-ABR: Soft Actor-Critic based deep reinforcement learning for Adaptive BitRate streaming.
Mandan NareshNandiraju GireeshParesh SaxenaManik GuptaPublished in: COMSNETS (2022)
Keyphrases
- actor critic
- reinforcement learning
- bit rate
- rate adaptation
- temporal difference
- policy gradient
- reinforcement learning algorithms
- video coding
- scalable video
- approximate dynamic programming
- bitstream
- neuro fuzzy
- video quality
- optimal control
- function approximation
- rate distortion
- gradient method
- model free
- motion vectors
- video streaming
- subband
- policy iteration
- state space
- multiview video coding
- markov decision processes
- recursive least squares
- image quality
- computational complexity
- rl algorithms
- learning algorithm
- adaptive control
- evaluation function
- optimal policy
- average reward
- temporal difference learning
- learning problems
- monte carlo
- multi agent
- machine learning
- action selection
- step size