FA-INR: Adaptive Implicit Neural Representations for Interpretable Exploration of Simulation Ensembles
This work addresses the problem of interpretable and efficient surrogate modeling for scientists analyzing complex simulation data, representing an incremental improvement over existing INR-based methods.
The paper tackles the challenge of efficiently exploring large-scale simulation ensembles by developing FA-INR, an adaptive implicit neural representation model that uses cross-attention and a mixture of experts to improve fidelity and interpretability, achieving high-fidelity results and enabling localized parameter-space exploration.
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.