Helpful Information Sharing for Partially Informed Planning Agents
Authors: Sarah Keren, David Wies, Sara Bernardini
IJCAI 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | 6 Empirical Evaluation The objective of the evaluation is to examine our approaches in terms of the computational resources used and their ability to find helpful interventions in a variety of domains. ... We compare three HIS solution approaches: BFS, Lazy-BFS (Section 4.1), and the Tka transformation (Section 4.2). ... Table 1 compares the computational resources used by BFS, Lazy-BFS and Tka. |
| Researcher Affiliation | Academia | 1The Taub Faculty of Computer Science, Technion Israel Institute of Technology 2 Benin School of Computer Science and Engineering, The Hebrew University of Jerusalem 3Department of Computer Science, Royal Holloway University of London |
| Pseudocode | No | A full description of Lazy BFS is given in the appendix1. |
| Open Source Code | Yes | 1Appendix: https://github.com/sarah-keren/HIS-IJCAI-23 |
| Open Datasets | Yes | We use seven partially observable planning domains: WUMPUS, TRAIL, COLOR-BALLS, COLORBALLS-E, LOGISTICS, and UNIX [Bonet and Geffner, 2011; Albore et al., 2009]. The Appendix includes a description of all the domains and the complete dataset and code. |
| Dataset Splits | No | The paper states 'For each domain, we generated at least 100 benchmarks by randomly selecting the initial state and the set of items that are known to the helper' but does not specify how the data is split into train, validation, and test sets for the experiments. |
| Hardware Specification | Yes | All evaluations were run on an 11th Gen Intel Core i9-11900F @ 2.50GHz 16. |
| Software Dependencies | No | The paper mentions software like PDDL, pyperplan, K-planner, Fast-Downward, and Lm-cut heuristic, but does not provide specific version numbers for any of them. |
| Experiment Setup | Yes | We run each instance on the three approaches with a time limit of 10 minutes, a memory limit of 2548 MB, and 1000 explored nodes. |