Selective Norm Monitoring

Authors: Natalia Criado, Jose M. Such

IJCAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental This section compares experimentally a NM with full prediction and a NM with approximate prediction to a traditional norm monitor. Extensive Simulation. We implemented a simulator in Java in which there is a set of agents that perform actions in a monitored environment described below. We conducted experiments in which the number of agents G took a random value within the J1, 100K interval.
Researcher Affiliation Academia Natalia Criado King s College London, UK natalia.criado@kcl.ac.uk Jose M. Such Lancaster University, UK j.such@lancaster.ac.uk
Pseudocode No No pseudocode or algorithm blocks are provided in the paper.
Open Source Code No No statement regarding the public release of source code or a link to a code repository is provided.
Open Datasets Yes We considered a real data set of patient confidentiality laws in the US, which are state-specific laws that forbid health departments to release personally identifiable information for specific causes when patients have communicable diseases. Confidentiality laws in the US7 cover 51 states with an average of over 57 regulations per state (for a total over 2900 laws). 7http://lawatlas.org/query?dataset=public-healthdepartments-and-state-patient-confidentiality-laws
Dataset Splits No The paper describes simulation experiments but does not explicitly mention standard training, validation, or test dataset splits in the context of model evaluation or reproducibility. It focuses on varying observation ratios and measuring detected violations over simulation steps.
Hardware Specification No The paper does not provide specific details about the hardware used for running the experiments or simulations.
Software Dependencies No The paper only mentions 'implemented a simulator in Java' without specifying the Java version or any other software dependencies with version numbers.
Experiment Setup Yes We implemented a simulator in Java in which there is a set of agents that perform actions in a monitored environment described below. We conducted experiments in which the number of agents G took a random value within the J1, 100K interval... The number of actions A took a random value within the J1, 20K interval. The simulation is executed 100 steps. We modelled different agent types with different capabilities... a set of roles is created at the beginning of each simulation. The number of roles created took a random value within the J1, AK interval.