Cargando…
Optimal Behavior is Easier to Learn than the Truth
We consider a reinforcement learning setting where the learner is given a set of possible models containing the true model. While there are algorithms that are able to successfully learn optimal behavior in this setting, they do so without trying to identify the underlying true model. Indeed, we sho...
| Autor principal: | |
|---|---|
| Formato: | Online Artículo Texto |
| Lenguaje: | English |
| Publicado: |
Springer Netherlands
2016
|
| Materias: | |
| Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5018263/ https://www.ncbi.nlm.nih.gov/pubmed/27682861 http://dx.doi.org/10.1007/s11023-016-9389-y |
| Sumario: | We consider a reinforcement learning setting where the learner is given a set of possible models containing the true model. While there are algorithms that are able to successfully learn optimal behavior in this setting, they do so without trying to identify the underlying true model. Indeed, we show that there are cases in which the attempt to find the true model is doomed to failure. |
|---|