NEMSim improves multi-event physical system simulations with rule integration
NEMSim: Learning Control-Conditioned Multi-Event Physical Dynamics via Executable Event-Mechanism Priors
Machine Learning
Summary
Simulating physical systems that have many interacting events controlled by different settings is usually very complex and slow. The authors created NEMSim, a tool that uses known rules about events to make simulation faster and more accurate, even when training data is limited. They also built a new 3D benchmark to measure performance. Their method outperforms previous approaches by a significant margin and works well even with less data.
What this means in practice
- •For simulation engineers: Generate faster and more accurate simulations of complex physical processes controlled by multiple inputs using rule-integrated neural models.
- •For game developers: Create more realistic physics-based behavior of complex systems in games by integrating known event rules into simulation engines.
Authors
Junsong Yu, Junjie Xie, Pengwei Liu, Dong Ni
Abstract
High-fidelity simulation of control-conditioned multi-event physical systems is computationally expensive, especially across broad control spaces and long trajectories. In these systems, macroscopic evolution emerges from localized discrete events whose intensities and effects depend on process controls and evolving local states, while the available system knowledge is typically expressed as event-attribute descriptions. Purely data-driven surrogates must infer these event effects from limited trajectory coverage, which can hinder generalization to unseen control regimes. Physics-guided methods instead primarily build on equation-level constraints or differentiable solvers rather than discrete event-rule priors. We therefore propose NEMSim (Neural Event-Mechanism Simulator), which compiles predefined event-attribute descriptions into an executable transition structure linking control-dependent event intensities, prior-guided mechanism attribution, and state-dependent responses. To enable evaluation of control-conditioned multi-event dynamics with explicit system knowledge, we construct a 3D KMC-based benchmark pairing high-fidelity trajectories with explicit event rules, standardized splits, and evaluation protocols. Across three settings, NEMSim reduces Avg. RMSE by 58.9%-81.3% relative to the strongest baseline in each setting. It also remains best in the data-efficiency study with training-data fractions down to 10%. Mechanism analyses further show that these gains arise from executable rule integration rather than prior access or architecture alone.