Crowdsourced Action-Model Acquisition for Planning
Authors: Hankz Zhuo
AAAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We finally exhibit that our approach is effective in the experiment. |
| Researcher Affiliation | Academia | Hankz Hankui Zhuo School of Advanced Computing, Sun Yat-sen University, Guangzhou, China. 510006 zhuohank@mail.sysu.edu.cn |
| Pseudocode | Yes | We show an overview of CAMA in Algorithm 1. |
| Open Source Code | No | The paper does not provide concrete access to source code (e.g., a repository link or an explicit code release statement) for the methodology described. |
| Open Datasets | Yes | We test our learning algorithm CAMA in three domains blocks4, driverlog5 and laundry. laundry is a domain created based on the description in the MIT PLIA1 dataset6. 4http://www.cs.toronto.edu/aips2000/aips-2000datafiles.tgz 5http://planning.cis.strath.ac.uk/competition/ 6http://architecture.mit.edu/house n/data/Place Lab/PLIA1.htm |
| Dataset Splits | No | The paper mentions collecting '75 plan traces as training data' and evaluates error rates, but does not specify distinct training, validation, and test splits with percentages or sample counts for reproducibility. |
| Hardware Specification | No | The paper does not provide any specific hardware details (e.g., CPU, GPU models, or memory specifications) used for running its experiments. |
| Software Dependencies | No | The paper mentions using a 'weighted MAX-SAT solver' but does not provide specific version numbers for this or any other software dependencies, which is necessary for reproducibility. |
| Experiment Setup | Yes | We varied the value of the parameter γ from 0 to 1 to see the trend of the accuracy, by fixing the number of plan traces to 60. Each time we randomly selected one of each five sequential intermediate partial states being observed in plan traces leaving other states empty, and each partial state was selected by 50% of propositions in the corresponding full state (a state was assumed to be represented by a set of propositions). |