Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].
Psychologically Based Virtual-Suspect for Interrogative Interview Training
Authors: Moshe Bitan, Galit Nahari, Zvi Nisin, Ariel Roth, Sarit Kraus
AAAI 2017 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments with 24 subjects demonstrate that the Virtual-Suspect s behavior is similar to that of a human who plays the role of the suspect. An experiment was conducted comparing the system s responding mechanism with that of a human instructor. |
| Researcher Affiliation | Academia | Moshe Bitan Department of Computer Science Bar-Ilan University, Israel EMAIL Galit Nahari Department of Criminology Bar-Ilan University, Israel EMAIL Zvi Nisin EMAIL Ariel Roth Department of Computer Science Bar-Ilan University, Israel EMAIL Sarit Kraus Department of Computer Science Bar-Ilan University, Israel EMAIL |
| Pseudocode | No | The paper contains figures illustrating processes (Figure 2: Response Model, Figure 3: Response Selection) and mathematical equations, but no explicitly labeled pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not contain any explicit statements about releasing source code, nor does it provide links to a code repository. |
| Open Datasets | No | The interrogation scenario chosen for the experiment is based on an actual burglary case from early 2013. The second interrogation scenario is also based on a real case from late 2014. These refer to real cases but no public dataset or access information is provided. |
| Dataset Splits | No | The paper describes experiments involving human subjects evaluating simulation transcripts, not the use of training, validation, or test dataset splits for machine learning model development or evaluation in the conventional sense. |
| Hardware Specification | No | The paper does not provide specific details about the hardware used to run the experiments, such as CPU, GPU models, or memory specifications. |
| Software Dependencies | No | The paper does not list any specific software dependencies with version numbers (e.g., Python, PyTorch, or specific solvers). |
| Experiment Setup | Yes | In addition, the Virtual-Suspect s personality profile was chosen to represent a moderately calm individual. More specifically, the initial internal-state vector was set to s0 = (0, 0, 3). Prior to the experiment, three different Virtual-Suspect models were simulated. In the first, a human instructor acted as the suspect. The second was the Virtual-Suspect response model, i.e. RMVS. In the third, a baseline randomized response selection mechanism was used. The three participants were informed that the interrogation will terminate by eliciting a confession. However, if a confession could not been reached, the simulation will terminate after 30 minutes. |