Schema-Guided Culture-Aware Complex Event Simulation with Multi-Agent Role-Play

TL;DR

This work develops a controllable complex news event simulator guided by both the event schema representing domain knowledge about the scenario and user-provided assumptions representing case-specific conditions and introduces a geo-diverse commonsense and cultural norm-aware knowledge enhancement component.

Abstract

Complex news events, such as natural disasters and socio-political conflicts, require swift responses from the government and society. Relying on historical events to project the future is insufficient as such events are sparse and do not cover all possible conditions and nuanced situations. Simulation of these complex events can help better prepare and reduce the negative impact. We develop a controllable complex news event simulator guided by both the event schema representing domain knowledge about the scenario and user-provided assumptions representing case-specific conditions. As event dynamics depend on the fine-grained social and cultural context, we further introduce a geo-diverse commonsense and cultural norm-aware knowledge enhancement component. To enhance the coherence of the simulation, apart from the global timeline of events, we take an agent-based approach to simulate the individual character states, plans, and actions. By incorporating the schema and cultural norms, our generated simulations achieve much higher coherence and appropriateness and are received favorably by participants from a humanitarian assistance organization.

Figures (5)

• Figure 1: A Simplified Overview of our proposed Miriam System for Complex Event Simulation.
• Figure 2: Figure depicting the global and character-level controllers during our simulation generation.
• Figure 3: In the event lifecycle, the global controller and each one of the characters plans its own events for the next time step. Then all of the plans are centralized and executed in temporal order. If the executed event involves other characters, the other character will be informed and replan its actions.
• Figure 4: Screengrab of the Miriam interface showing an example simulation for a disease outbreak in Indonesia. The simulation is visualized in the form of an event timeline, with each event provided with a detailed description including related socio-cultural norms, along with background details of the characters involved in the event.
• Figure 5: Results from utility evaluation by participants from a humanitarian assistance organization.