FA-INR: Adaptive Implicit Neural Representations for Interpretable Exploration of Simulation Ensembles

TL;DR

FA-INR introduces an adaptive, interpretable surrogate modeling approach for ensemble simulations using cross-attention and mixture of experts, enabling localized exploration and scientific insights.

Contribution

It proposes a novel adaptive INR framework with cross-attention and MoE that improves scalability, interpretability, and localized exploration of complex simulation data.

Findings

FA-INR effectively models complex localized structures in simulation ensembles.

The learned experts produce interpretable partitions revealing scientific insights.

FA-INR supports localized sensitivity analysis and domain exploration.

Abstract

Surrogate models are essential for efficient exploration of large-scale ensemble simulations. Implicit neural representations (INRs) provide a compact and continuous framework for modeling spatially structured data, but they often struggle with learning complex localized structures within the scientific fields. Recent INR-based surrogates address this by augmenting INRs with explicit feature structures, but at the cost of flexibility and substantial memory overhead. In this paper, we present Feature-Adaptive INR (FA-INR), an adaptive INR-based surrogate model for high-fidelity and interpretable exploration of ensemble simulations. Instead of relying on structured feature representations, FA-INR leverages cross-attention over a learnable key-value memory bank to allocate model capacity adaptively based on the data characteristics. To further improve scalability, we introduce a coordinate-guided mixture of experts (MoE) framework that enhances both efficiency and specialization of feature representations. More importantly, the learned experts produce an interpretable partition over the simulation domain, enabling scientists to identify complex structures and perform localized parameter-space exploration. Beyond quantitative and qualitative evaluations, we also demonstrate that our learned expert specialization can reveal meaningful scientific insights and support localized sensitivity analysis.