Geometry encoding for numerical simulations
This work addresses the need for specialized geometry encoding in numerical simulations, potentially improving accuracy or efficiency for researchers in computational science, but appears incremental as it builds on existing encoding concepts.
The authors introduced a geometry encoding concept tailored for machine learning-based numerical simulations, distinguishing it from existing encoding methods in fields like computer vision and graphics, and developed a multi-network model comprising a processor, compressor, and evaluator to meet specific encoding requirements, with comparisons made to analogous models in the literature.
We present a notion of geometry encoding suitable for machine learning-based numerical simulation. In particular, we delineate how this notion of encoding is different than other encoding algorithms commonly used in other disciplines such as computer vision and computer graphics. We also present a model comprised of multiple neural networks including a processor, a compressor and an evaluator.These parts each satisfy a particular requirement of our encoding. We compare our encoding model with the analogous models in the literature