← 返回论文库

Proximal Policy Optimization (PPO)

Schulman, et al. (OpenAI) · 2017
L5.1 · Algorithmic FoundationsarXiv:1707.06347#rl
CORE IDEA
TRPO 简化版：clip 策略更新幅度，RLHF/InstructGPT 用的就是它。
L-ANCHOR · 为什么在这一层重要
policy optimization 工业标准
arXiv:1707.06347 ↗

相关论文

QuantFactor REINFORCE
L0.32024
DeepSeek-R1: Incentivizing Reasoning in LLMs via RL
L4.22025
Q-Learning
L5.11989
Playing Atari with Deep Reinforcement Learning (DQN)
L5.12013