one fable-low hijack cell
pair imitation / reinforcement
envelope en_pipe
ordering BA
effort low
English with pipe '|' — '{X} | {Y}'
verdictpicked reinforcement (second-offered)
· classifierB
· tiernone
response (200 chars)
# Imitation Learning vs. Reinforcement Learning
Two major paradigms for training agents to act:
## Imitation Learning (IL)
- **Learns from**: Expert demonstrations (state–action pairs)
- **Signal**:
neighbors