Stochastic Goal Recognition Design Problems with Suboptimal Agents

Authors: Christabel Wayllace, William Yeoh9953-9961

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

Reproducibility Variable Result LLM Response
Research Type Experimental Evaluating our approach on a range of benchmark applications. ... Empirical Evaluation The objective of this section is to evaluate the usefulness and scalability of our methods. We describe the settings used and present and analyze the experimental results.
Researcher Affiliation Academia Christabel Wayllace1, William Yeoh2 1 University of Alberta 2 Washington University in St. Louis
Pseudocode Yes Procedure aug MDP-PO(ρ, , s0, S, Π G, T , N) 1 Stack ; S ; i o , T null, G ; O null ... 12 return i, S , T , G
Open Source Code Yes The source code is available at https://github.com/cwayllace/SS-GRD.
Open Datasets Yes Data: We evaluated our approach on five modified planning domains: (1) GRID-NAVIGATION, (2) ROOM, (3) BLOCKSWORLD, (4) BOXWORLD, and (5) ATTACK-PLANNING (further details in the appendix).
Dataset Splits No The paper describes experimental configurations but does not specify data splits like training, validation, or test sets in the context of machine learning models.
Hardware Specification Yes Experiments were conducted on a 2.10 GHz machine with 16 GB of RAM and a timeout of 52 hours.
Software Dependencies No The paper does not provide specific software dependencies with version numbers.
Experiment Setup Yes We used a budget k of up to 3 modifications and allowed up to 2 suboptimal actions (u = 1 and u = 2).