Figure 6 - ICU-Sepsis: A Benchmark MDP Built from Real Medical Data

Figure 6: Effects of removing some actions from the set of admissible actions on the learned policies as the probability of removing actions (σ) increases from 0 to 1. Each perturbation was done 32 times for each environment and the average and standard error of the results are shown. (a) The average return for different policies. (b) The average lengths of episodes for different policies.

Effects of removing some actions from the set of admissible actions on the learned policies as the probability of removing actions (σ) increases from 0 to 1. Each perturbation was done 32 times for each environment and the average and standard error of the results are shown. (a) The average return for different policies. (b) The average lengths of episodes for different policies.

Go to figure page

Reference

ICU-Sepsis: A Benchmark MDP Built from Real Medical Data - Scientific Figure on ResearchGate. Available from: https://www.researchgate.net/figure/Effects-of-removing-some-actions-from-the-set-of-admissible-actions-on-the-learned_fig5_381307915 [accessed 19 May 2025]

Caption

Figure 6: Effects of removing some actions from the set of admissible actions on the learned policies as the probability of removing actions (σ) increases from 0 to 1. Each perturbation was done 32 times for each environment and the average and standard error of the results are shown. (a) The average return for different policies. (b) The average lengths of episodes for different policies.

Embed code

<a href="https://www.researchgate.net/figure/Effects-of-removing-some-actions-from-the-set-of-admissible-actions-on-the-learned_fig5_381307915"><img src="https://www.researchgate.net/publication/381307915/figure/fig5/AS:11431281250974703@1718080067106/Effects-of-removing-some-actions-from-the-set-of-admissible-actions-on-the-learned.ppm" alt="Effects of removing some actions from the set of admissible actions on the learned policies as the probability of removing actions (σ) increases from 0 to 1. Each perturbation was done 32 times for each environment and the average and standard error of the results are shown. (a) The average return for different policies. (b) The average lengths of episodes for different policies."/></a>