Skip to main content
Cross Validated

Return to Answer

up-to-date link to the Dearden et al paper and inline the link to Houthooft et al paper
Source Link
edit approved Jun 21 at 7:50
cottontail
edit approved Jun 21 at 7:50
cottontail
  • 1.1k
  • 5
  • 9
  • 16

For small tasks you can use Bayesian Q-Learning https://ocs.aaai.org/Papers/AAAI/1998/AAAI98-108.pdfBayesian Q-Learning.

For larger tasks, you can use its variational version https://pdfs.semanticscholar.org/ce2e/0bd9135814f4018106bc31d87902b358e251.pdfvariational version.

The core idea is to systematically try suboptimal x which shall pay off in the long term.

For small tasks you can use Bayesian Q-Learning https://ocs.aaai.org/Papers/AAAI/1998/AAAI98-108.pdf

For larger tasks you can use its variational version https://pdfs.semanticscholar.org/ce2e/0bd9135814f4018106bc31d87902b358e251.pdf

The core idea is to systematically try suboptimal x which shall pay off in the long term.

For small tasks you can use Bayesian Q-Learning.

For larger tasks, you can use its variational version.

The core idea is to systematically try suboptimal x which shall pay off in the long term.

Source Link
Karel Macek
Karel Macek
  • 2.8k
  • 15
  • 26

For small tasks you can use Bayesian Q-Learning https://ocs.aaai.org/Papers/AAAI/1998/AAAI98-108.pdf

For larger tasks you can use its variational version https://pdfs.semanticscholar.org/ce2e/0bd9135814f4018106bc31d87902b358e251.pdf

The core idea is to systematically try suboptimal x which shall pay off in the long term.