Nov 12, 2021 · Title:Human irrationality: both bad and good for reward inference. Authors:Lawrence Chan, Andrew Critch, Anca Dragan. View a PDF of the paper ...
Human irrationality: both bad and good for reward inference · 20 Citations · 39 References.
Oct 16, 2020 · We find that incorrectly assuming noisy-rationality for an irrational demonstrator can lead to remarkably poor reward inference accuracy.
Feb 12, 2024 · Occam's razor is insufficient to infer the preferences of irrational agents · Human irrationality: both bad and good for reward inference. I ...
Co-authors ; Human irrationality: both bad and good for reward inference. L Chan, A Critch, A Dragan. arXiv preprint arXiv:2111.06956, 2021. 24, 2021.
Human irrationality: both bad and good for reward inference. L Chan, A Critch, A Dragan. arXiv preprint arXiv:2111.06956, 2021. 24, 2021. Optimal cost design ...
Each irrationality uses the parameter value that is most informative. As discussed in section 3.2, different irrational- ity types have different slopes and ...
While demonstrations do pro- vide the most information when the human is highly ratio- nal, comparisons gain an advantage when querying a more irrational human.
Sep 3, 2024 · Chan, L., Critch, A., and Dragan, A. Human irrationality: both bad and good for reward inference. arXiv preprint. arXiv:2111.06956, 2021. Chen, ...
Chan et al., Human Irrationality: Both Bad and Good for Reward Inference (2021). Julian et al., Never Stop Learning: The Effectiveness of Fine-Tuning in ...