SLALOM: Simulation Lifecycle Analysis via Longitudinal Observation Metrics for Social Simulation
SLALOM: Simulation Lifecycle Analysis via Longitudinal Observation Metrics for Social Simulation
Authors: Juhoon Lee, Joseph Seering Date: 2026-04-13 Paper ID: openalex:2604.11466
Summary
This paper addresses the validity crisis in LLM-based social simulation by moving from outcome-centric evaluation to process-based validation. The authors introduce SLALOM, a framework that employs Pattern-Oriented Modeling to define intermediate phase-based constraints (SLALOM gates) within social simulations. By applying Dynamic Time Warping to align simulated multivariate trajectories with empirical ground truth, the framework provides a quantitative mechanism to ensure structural realism and sociological plausibility in agent-based modeling.
Key Contributions
- Introduces SLALOM, a framework for evaluating LLM agent social simulations based on process fidelity rather than final outcome accuracy.
- Utilizes Pattern-Oriented Modeling to define intermediate ‘SLALOM gates’ as structural constraints for social simulation trajectories.
- Applies Dynamic Time Warping (DTW) to quantify alignment between simulated multivariate time series and empirical social science data.
Open Questions & Future Work
Key Concepts
- slalom-framework: A framework for evaluating social simulations by enforcing process fidelity through intermediate temporal constraints rather than focusing solely on final outcomes.
Archivist Review
I have approved the SLALOM framework as it establishes a distinct methodology for process-based evaluation of agent-based models. I also approved the open question regarding non-linear social simulations, as it identifies a clear, persistent boundary of current temporal alignment metrics like DTW. Other minor technical components were rejected to maintain the focus on the primary architectural innovation of the paper.
Approved Concepts
- SLALOM (Simulation Lifecycle Analysis via Longitudinal Observation Metrics): It provides a novel process-centric evaluation paradigm for agent-based social simulations, specifically addressing the validity crisis in LLM-based modeling.
Approved Open Questions
- Evaluating Non-Linear Social Simulations: This is a critical limitation for the scalability and applicability of process-based validation frameworks to diverse social phenomena where non-linearity is a core feature.
Links
Metadata & Links
- url
- https://arxiv.org/abs/2604.11466
- paper_id
- 2604.11466
- paper_source
- openalex
- domain
- nlp
- tags
-
- architectures
-
- datasets
-
- concept_slugs
- slalom-framework
- dataset_slugs
-
- skill
- TimeSeriesSkill
- created_at
- 2026-04-16T05:47:18Z