Admissible Policy Teaching through Reward Design

Authors: Kiarash Banihashem, Adish Singla, Jiarui Gan, Goran Radanovic6037-6045

AAAI 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We design a local search algorithm to solve the surrogate problem and showcase its utility using simulation-based experiments.
Researcher Affiliation Academia Max Planck Institute for Software Systems {kbanihas, adishs, jrgan, gradanovic}@mpi-sws.org
Pseudocode Yes Algorithm 1. CONSTRAIN&OPTIMIZE
Open Source Code No For details regarding the experiments and code, please refer to the full version of our paper (Banihashem et al. 2022).
Open Datasets No As an experimental testbed, we consider three simple navigation environments, shown in Figure 2.
Dataset Splits No The paper describes custom environments and mentions parameters, but does not specify explicit training, validation, or test dataset splits.
Hardware Specification No The paper does not provide specific hardware details used for running experiments.
Software Dependencies No The paper does not specify software names with version numbers.
Experiment Setup Yes By default, we set the parameters γ = 0.9, λ = 1.0 and ϵ = 0.1.