AATEAM: Achieving the Ad Hoc Teamwork by Employing the Attention Mechanism
Authors: Shuo Chen, Ewa Andrejczuk, Zhiguang Cao, Jie Zhang7095-7102
AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We perform extensive experiments to show the effectiveness of our method. [...] In this section, we present the experiments we have done to demonstrate the effectiveness of AATEAM. |
| Researcher Affiliation | Collaboration | 1School of Computer Science and Engineering, Nanyang Technological University, Singapore 2ST Engineering NTU Corporate Laboratory, Nanyang Technological University, Singapore 3Department of Industrial Systems Engineering and Management, National University of Singapore, Singapore |
| Pseudocode | No | The paper describes its method through figures and textual explanations but does not include a formally labeled “Pseudocode” or “Algorithm” block. |
| Open Source Code | No | The paper does not contain an explicit statement about releasing its source code or provide a link to a code repository for the AATEAM methodology. |
| Open Datasets | Yes | Following Barrett et al. (2017), we use a limited version of the Robo Cup 2D domain, i.e. half field offense (HFO) domain (Hausknecht et al. 2016). [...] We run the binary file released from the 2013 Robo Cup 2D simulation league competition (Robo Cup 2013). |
| Dataset Splits | No | The paper describes data collection for training and testing, but does not explicitly mention the use of a separate validation set or provide details on its split or purpose for hyperparameter tuning. |
| Hardware Specification | No | The paper mentions a “decision time limit” but does not specify the exact hardware components (e.g., CPU, GPU models, memory) used for running the experiments or training the models. |
| Software Dependencies | No | The paper refers to algorithms and optimizers used (e.g., “SARSA algorithm”, “Adam optimizer”, “GRU”) and mentions “Relu function” and “Tanh function”, but does not provide specific version numbers for any software libraries or dependencies used in the implementation. |
| Experiment Setup | Yes | We sum up the experiment parameters in Table 1. The number of hidden state layers for GRU is 5. [...] The rate of dropout is 0.1. When training the attention networks, the default learning rate is 0.01 and the batch size is 32. [...] Moreover, during the training, we apply the teacher forcing strategy (Williams and Zipser 1989), i.e. use ˆat instead of at as the input to the decoder in the step t + 1. |