Information Shaping for Enhanced Goal Recognition of Partially-Informed Agents
Authors: Sarah Keren, Haifeng Xu, Kofi Kwapong, David Parkes, Barbara Grosz9908-9915
AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We demonstrate the effectiveness and efficiency of the suggested method on standard benchmarks. In this section, we report experiments that demonstrate both the effect of sensor extensions on WCD as well as the computational advantage provided through CG-Pruning. Results. Tables 1 and 2 summarize the results for the FD and FF solvers, respectively. |
| Researcher Affiliation | Academia | Sarah Keren, Haifeng Xu, KofiKwapong, David Parkes, Barbara Grosz School of Engineering and Applied Sciences Harvard University {skeren, hxu}@seas.harvard.edu, kwapongk@college.harvard.edu, {parkes, grosz}@eecs.harvard.edu |
| Pseudocode | No | The paper describes algorithms and methods verbally and formally with definitions, but it does not include a block or figure explicitly labeled as "Pseudocode" or "Algorithm". |
| Open Source Code | Yes | A complete account of our dataset and code can be found in https://github.com/sarah-keren/GRD-APK-AAAI-20-Appendixand-Experiments |
| Open Datasets | Yes | Dataset. We use seven domains, adapted from Bonet and Geffner (2011) and Albore, Palacios, and Geffner (2009). The adaptation from partially observable planning to GRDAPK involves specifying for each instance the set of possible goals and sensor extensions. A complete account of our dataset and code can be found in https://github.com/sarah-keren/GRD-APK-AAAI-20-Appendixand-Experiments |
| Dataset Splits | No | The paper does not explicitly provide training, validation, and test dataset splits with specific percentages, counts, or references to predefined splits. |
| Hardware Specification | No | The paper does not provide any specific details about the hardware used for the experiments, such as GPU/CPU models, memory, or cloud instance types. |
| Software Dependencies | No | The paper mentions several software components like "k-planner", "FF classical planner", "Fast-Downward (FD)", "pyperplan", and "STRIPS", but it does not specify their version numbers. |
| Experiment Setup | Yes | We use 40 instances for each domain, using design budgets of 1 and 2. We fix the time limit for an execution to 20 minutes and 1000 search steps (each corresponding to a modification set), whichever was first. |