SysCaps: Language Interfaces for Simulation Surrogates of Complex Systems

TL;DR

This paper introduces SysCaps, a novel language interface for simulation surrogates of complex systems, enabling more accessible, accurate, and generalizable modeling through natural language descriptions and large language model synthesis.

Contribution

The work presents a new multimodal model and training pipeline that use language descriptions to improve surrogate model accuracy and generalization for complex energy systems.

Findings

SysCaps-augmented surrogates outperform traditional methods in accuracy.

Models exhibit improved generalization to semantically related descriptions.

Potential for language-driven design space exploration and regularization.

Abstract

Surrogate models are used to predict the behavior of complex energy systems that are too expensive to simulate with traditional numerical methods. Our work introduces the use of language descriptions, which we call ``system captions'' or SysCaps, to interface with such surrogates. We argue that interacting with surrogates through text, particularly natural language, makes these models more accessible for both experts and non-experts. We introduce a lightweight multimodal text and timeseries regression model and a training pipeline that uses large language models (LLMs) to synthesize high-quality captions from simulation metadata. Our experiments on two real-world simulators of buildings and wind farms show that our SysCaps-augmented surrogates have better accuracy on held-out systems than traditional methods while enjoying new generalization abilities, such as handling semantically related descriptions of the same test system. Additional experiments also highlight the potential of SysCaps to unlock language-driven design space exploration and to regularize training through prompt augmentation.