Proximal Policy Optimization Algorithms
At a glance
PPO improves an agent's behavior through repeated updates guided by trial and reward.
In one paragraph
Introduced PPO, a simpler method for learning through trial and reward, using repeated updates from collected experience to improve performance on games and simulated movement.
Authors 5
- John Schulman
- Filip Wolski
- Prafulla Dhariwal
- Alec Radford
- Oleg Klimov
