Cognitive Social Learners: An Architecture for Modeling Normative Behavior
Authors: Rahmatollah Beheshti, Awrad Mohammed Ali, Gita Sukthankar
AAAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments To demonstrate the utility of our normative architecture at modeling the adoption of sustainable practices, two case studies are presented. In first case study, we evaluate the performance of CSL at simulating norm emergence in a park scenario, as compared to the normative BDI (NBDI) and social learning (SL) architectures. The second case study is designed to evaluate the ability of CSL to model the propagation of norms in real-world environments. We compare the performance of our proposed architecture with an existing architecture for simulating the propagation of smoking norms.Results Our proposed framework (CSL) was compared against two other benchmarks. The first one, NBDI, is a version of the normative BDI architecture described in Criado, Argente, and Botti (2010), and the second one, SL, is the social learning framework introduced in Sen and Airiau (2007). In order to make a fair comparison between different architectures, the NBDI and SL frameworks are implemented by removing some of the components of CSL. |
| Researcher Affiliation | Academia | Rahmatollah Beheshti Department of EECS University of Central Florida Orlando, FL beheshti@eecs.ucf.edu Awrad Mohammed Ali Department of EECS University of Central Florida Orlando, FL awrad.emad@knights.ucf.edu Gita Sukthankar Department of EECS University of Central Florida Orlando, FL gitars@eecs.ucf.edu |
| Pseudocode | Yes | Figure 2: CSL pseudocode (blf=Beliefs, des=Desires, pln=Plans, rew=Rewards, san=Sanctions) |
| Open Source Code | No | The paper does not provide any statement or link indicating that the source code for the methodology is openly available. |
| Open Datasets | No | The paper describes scenarios (e.g., 'littering scenario', 'smoke-free campus initiative') and uses phrases like 'real-world data' for the smoking cessation study, but does not provide concrete access information (link, DOI, specific citation with author/year, or recognized public dataset name with attribution) for any datasets used. |
| Dataset Splits | No | The paper describes simulation runs ('average of 20 runs') and population size ('1000 agents'), but it does not specify explicit training, validation, or test dataset splits in terms of percentages, sample counts, or defined subsets for model reproduction. |
| Hardware Specification | No | The paper does not specify any particular hardware components (e.g., CPU, GPU models, memory) used for running the experiments. |
| Software Dependencies | No | The paper mentions implementing 'Q-learning algorithm' but does not specify any software dependencies with version numbers (e.g., programming languages, libraries, frameworks, or solvers with specific versions). |
| Experiment Setup | Yes | For these experiments, we fixed the population size at 1000. There is an observable vicinity defined for each agent. Within that range an agent can observe other agents actions. A certain percentage of agents are assumed to be punishers (20 percent)... The certainty values (δ) for beliefs and desires are assigned uniformly at random at the beginning of the scenario. |