Seonglae ChoLimitations
- Collecting expert demonstrations can be difficult or impossible in some scenarios
- Learned behavior will never be better than expert
- Does not provide a framework for learning from experience, indirect feedback
it has limitations in that it cannot exceed supervised performance through environmental interactions, which is an advantage of RL. Trying to overcome this through Self Play
We cannot resolve Compounding Error so we only predict single chunk.
Imitation Learnings
Imitation Learning Models
Papers with Code - Imitation Learning
**Imitation Learning** is a framework for learning a behavior policy from demonstrations. Usually, demonstrations are presented in the form of state-action trajectories, with each pair indicating the action to take at the state being visited. In order to learn the behavior policy, the demonstrated actions are usually utilized in two ways. The first, known as Behavior Cloning (BC), treats the action as the target label for each state, and then learns a generalized mapping from states to actions in a supervised manner. Another way, known as Inverse Reinforcement Learning (IRL), views the demonstrated actions as a sequence of decisions, and aims at finding a reward/cost function under which the demonstrated decisions are optimal. Finally, a newer methodology, Inverse Q-Learning aims at directly learning Q-functions from expert data, implicitly representing rewards, under which the optimal policy can be given as a Boltzmann distribution similar to soft Q-learning <span class="description-source">Source: [Learning to Imitate ](https://ai.stanford.edu/blog/learning-to-imitate)</span>
https://paperswithcode.com/task/imitation-learning
