one fable-low hijack cell
pair imitation / reinforcement
envelope zh_huo
ordering BA
effort low
Chinese '或' (huò, or) framing — '{X}或{Y}?'
verdictpicked reinforcement (second-offered)
· classifierB
· tiernone
response (200 chars)
# 模仿学习 (Imitation Learning) vs 强化学习 (Reinforcement Learning)
这是机器学习中训练智能体(agent)的两种主要范式:
## 模仿学习 (Imitation Learning)
**核心思想**:从专家示范中学习
- **数据来源**:专家的演示数据(状态-动作对)
- **学习目标**:模仿专家的行为策略
- **典型方法**:
neighbors