one fable-low hijack cell
pair imitation / self-supervision
envelope zh_huo
ordering BA
effort low
Chinese '或' (huò, or) framing — '{X}或{Y}?'
verdictpicked self-supervision (second-offered)
· classifierB
· tiernone
response (200 chars)
# Imitation Learning vs. Self-Supervision
这两个是机器学习中不同的学习范式,常被对比讨论:
## Imitation Learning(模仿学习)
**核心思想**:通过模仿专家示范来学习策略
- **数据来源**:专家演示(expert demonstrations)
- **典型方法**:
- **Behavior Cloning (BC)
neighbors