Computer Science – Learning
Scientific paper
2010-11-02
Computer Science
Learning
Appearing in the 14th International Conference on Artificial Intelligence and Statistics (AISTATS 2011)
Scientific paper
Sequential prediction problems such as imitation learning, where future observations depend on previous predictions (actions), violate the common i.i.d. assumptions made in statistical learning. This leads to poor performance in theory and often in practice. Some recent approaches provide stronger guarantees in this setting, but remain somewhat unsatisfactory as they train either non-stationary or stochastic policies and require a large number of iterations. In this paper, we propose a new iterative algorithm, which trains a stationary deterministic policy, that can be seen as a no regret algorithm in an online learning setting. We show that any such no regret algorithm, combined with additional reduction assumptions, must find a policy with good performance under the distribution of observations it induces in such sequential settings. We demonstrate that this new approach outperforms previous approaches on two challenging imitation learning problems and a benchmark sequence labeling problem.
Bagnell Andrew J.
Gordon Geoffrey J.
Ross Stephane
No associations
LandOfFree
A Reduction of Imitation Learning and Structured Prediction to No-Regret Online Learning does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with A Reduction of Imitation Learning and Structured Prediction to No-Regret Online Learning, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and A Reduction of Imitation Learning and Structured Prediction to No-Regret Online Learning will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-595968