A Dantzig Selector Approach to Temporal Difference Learning

@inproceedings{Geist2012ADS,
  title={A Dantzig Selector Approach to Temporal Difference Learning},
  author={Matthieu Geist and Bruno Scherrer and Alessandro Lazaric and Mohammad Ghavamzadeh},
  booktitle={ICML},
  year={2012}
}
LSTD is a popular algorithm for value function approximation. Whenever the number of features is larger than the number of samples, it must be paired with some form of regularization. In particular, `1-regularization methods tend to perform feature selection by promoting sparsity, and thus, are wellsuited for high–dimensional problems. However, since LSTD… CONTINUE READING