VicSim: Enhancing Victim Simulation with Emotional and Linguistic Fidelity

2025-01-06

AI Summary

Current public safety training suffers from low-fidelity victim simulation, characterized by distorted emotional expression and flattened linguistic style, undermining training effectiveness. To address this, we propose a high-fidelity victim simulation system that integrates scenario-aware victim modeling with a generative adversarial network (GAN) framework—its first such application in this domain. We innovatively introduce a key-information-guided prompting mechanism and design the discriminator to prioritize syntactic features and fine-grained emotional dynamics. The system synergistically combines GANs with large language models (LLMs) to enable personalized language generation and temporally coherent emotional evolution modeling. Multidimensional human evaluation demonstrates statistically significant improvements over GPT-4 across three core metrics: emotional authenticity, linguistic naturalness, and human-machine similarity. This work establishes a verifiable, high-fidelity paradigm for interactive emergency response training.

Technology Category

• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior
• Natural Language Processing: Generation
• Computer Vision: Generative Adversarial Networks (GANs) for Vision

Application Category

• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Scenario-based training has been widely adopted in many public service sectors. Recent advancements in Large Language Models (LLMs) have shown promise in simulating diverse personas to create these training scenarios. However, little is known about how LLMs can be developed to simulate victims for scenario-based training purposes. In this paper, we introduce VicSim (victim simulator), a novel model that addresses three key dimensions of user simulation: informational faithfulness, emotional dynamics, and language style (e.g., grammar usage). We pioneer the integration of scenario-based victim modeling with GAN-based training workflow and key-information-based prompting, aiming to enhance the realism of simulated victims. Our adversarial training approach teaches the discriminator to recognize grammar and emotional cues as reliable indicators of synthetic content. According to evaluations by human raters, the VicSim model outperforms GPT-4 in terms of human-likeness.

Problem

• Emotional Expression
• Language Simulation
• Realism in Simulation

Innovation

• Generative Adversarial Network (GAN)
• Critical Information Prompting
• Emulation Fidelity Enhancement