Adapting Plans through Communication with Unknown Teammates
Authors: Trevor Sarratt
AAAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In our submission to AAMAS-16 (Sarratt and Jhala Pending), we provided an efficient procedure for evaluating potential state-action pairs for communication by examining the uncertainty within a teammate model as well as empirically analyzed various aspects of the communicative capability. For our test domain, we used a variation of the multiagent pursuit domain where two agents attempt to capture a prey within a maze. We were able to show the tradeoff between collected information and communication rates and, likewise, the effect of communication costs on query rates and the resulting expected utility of the agent. Finally, our analysis determined that branch points in a maze are commonly communicated more frequently than neighboring cells, indicating that the tested agents determined resolving uncertainty at such states was associated with higher utility. |
| Researcher Affiliation | Academia | Trevor Sarratt University of California Santa Cruz tsarratt@soe.ucsc.edu |
| Pseudocode | No | The paper does not contain any clearly labeled 'Pseudocode' or 'Algorithm' blocks, nor does it present structured steps in a code-like format. |
| Open Source Code | No | The paper does not include any explicit statements about making the source code available, nor does it provide a link to a code repository. |
| Open Datasets | No | The paper states, 'For our test domain, we used a variation of the multiagent pursuit domain where two agents attempt to capture a prey within a maze.' This describes a custom test domain, not a publicly available or open dataset, and no citation or access information is provided. |
| Dataset Splits | No | The paper does not specify exact percentages or sample counts for training, validation, or test splits. It refers to a 'test domain' but not a formal data split. |
| Hardware Specification | No | The paper does not provide any specific hardware details such as GPU models, CPU types, or memory specifications used for running the experiments. |
| Software Dependencies | No | The paper does not provide specific software names with version numbers that would be necessary to replicate the experiment. |
| Experiment Setup | No | The paper does not provide specific details about the experimental setup, such as hyperparameter values (e.g., learning rate, batch size) or other system-level training settings. |