arXiv subjects

Samundra Karki

Publications and source records attributed to Samundra Karki.

5 recordsLinked to original sources

arXiv2026

GENIE: Gram-Eigenmode INR Editing with Closed-Form Geometry Updates

Implicit Neural Representations (INRs) provide compact models of geometry, but it is unclear when their learned shapes can be edited without retraining. We show that the Gram operator induced by the INR's penultimate features admits deformation eigenmodes that parameterize a family of realizable edits of the SDF zero level set. A key finding is that these modes are not intrinsic to the geometry alone: they are reliably recoverable only when the Gram operator is estimated from sufficiently rich sampling distributions. We derive a single closed-form update that performs geometric edits to the INR without optimization by leveraging the deformation modes. We characterize theoretically the precise set of deformations that are feasible under this one-shot update, and show that editing is well-posed exactly within the span of these deformation modes.

cs.GR↗

arXiv2026

Neural Geometry for PDEs: Regularity, Stability, and Convergence Guarantees

Implicit Neural Representations (INRs) have emerged as a powerful tool for geometric representation, yet their suitability for physics-based simulation remains underexplored. While metrics like Hausdorff distance quantify surface reconstruction quality, they fail to capture the geometric regularity required for provable numerical performance. This work establishes a unified theoretical framework connecting INR training errors to Partial Differential Equation (PDE) (specifically, linear elliptic equation) solution accuracy. We define the minimal geometric regularity required for INRs to support well-posed boundary value problems and derive \emph{a priori} error estimates linking the neural network's function approximation error to the finite element discretization error. Our analysis reveals that to match the convergence rate of linear finite elements, the INR training loss must scale quadratically relative to the mesh size.

cs.CE↗

arXiv2025

Mechanics Simulation with Implicit Neural Representations of Complex Geometries

Implicit Neural Representations (INRs), characterized by neural network-encoded signed distance fields, provide a powerful means to represent complex geometries continuously and efficiently. While successful in computer vision and generative modeling, integrating INRs into computational analysis workflows, such as finite element simulations, remains underdeveloped. In this work, we propose a computational framework that seamlessly combines INRs with the Shifted Boundary Method (SBM) for high-fidelity linear elasticity simulations without explicit geometry transformations. By directly querying the neural implicit geometry, we obtain the surrogate boundaries and distance vectors essential for SBM, effectively eliminating the meshing step. We demonstrate the efficacy and robustness of our approach through elasticity simulations on complex geometries (Stanford Bunny, Eiffel Tower, gyroids) sourced from triangle soups and point clouds. Our method showcases significant computational advantages and accuracy, underscoring its potential in biomedical, geophysical, and advanced manufacturing applications.

cs.CE↗

arXiv2025

Direct Flow Simulations with Implicit Neural Representation of Complex Geometry

Implicit neural representations have emerged as a powerful approach for encoding complex geometries as continuous functions. These implicit models are widely used in computer vision and 3D content creation, but their integration into scientific computing workflows, such as finite element or finite volume simulations, remains limited. One reason is that conventional simulation pipelines require explicit geometric inputs (meshes), forcing INR-based shapes to be converted to meshes--a step that introduces approximation errors, computational overhead, and significant manual effort. Immersed boundary methods partially alleviate this issue by allowing simulations on background grids without body-fitted meshes. However, they still require an explicit boundary description and can suffer from numerical artifacts, such as sliver cut cells. The shifted boundary method (SBM) eliminates the need for explicit geometry by using grid-aligned surrogate boundaries, making it inherently compatible with implicit shape representations. Here, we present a framework that directly couples neural implicit geometries with SBM to perform high-fidelity fluid flow simulations without any intermediate mesh generation. By leveraging neural network inference, our approach computes the surrogate boundary and distance vectors required by SBM on-the-fly directly from the INR, thus completely bypassing traditional geometry processing. We demonstrate this approach on canonical 2D and 3D flow benchmarks (lid-driven cavity flows) and complex geometries (gyroids, the Stanford bunny, and AI-generated shapes), achieving simulation accuracy comparable to conventional mesh-based methods. This work highlights a novel pathway for integrating AI-driven geometric representations into computational physics, establishing INRs as a versatile and scalable tool for simulations and removing a long-standing bottleneck in geometry handling.

cs.GR↗

arXiv2024

FlowBench: A Large Scale Benchmark for Flow Simulation over Complex Geometries

Simulating fluid flow around arbitrary shapes is key to solving various engineering problems. However, simulating flow physics across complex geometries remains numerically challenging and computationally resource-intensive, particularly when using conventional PDE solvers. Machine learning methods offer attractive opportunities to create fast and adaptable PDE solvers. However, benchmark datasets to measure the performance of such methods are scarce, especially for flow physics across complex geometries. We introduce FlowBench, a dataset for neural simulators with over 10K samples, which is currently larger than any publicly available flow physics dataset. FlowBench contains flow simulation data across complex geometries (\textit{parametric vs. non-parametric}), spanning a range of flow conditions (\textit{Reynolds number and Grashoff number}), capturing a diverse array of flow phenomena (\textit{steady vs. transient; forced vs. free convection}), and for both 2D and 3D. FlowBench contains over 10K data samples, with each sample the outcome of a fully resolved, direct numerical simulation using a well-validated simulator framework designed for modeling transport phenomena in complex geometries. For each sample, we include velocity, pressure, and temperature field data at 3 different resolutions and several summary statistics features of engineering relevance (such as coefficients of lift and drag, and Nusselt numbers). %Additionally, we include masks and signed distance fields for each shape. We envision that FlowBench will enable evaluating the interplay between complex geometry, coupled flow phenomena, and data sufficiency on the performance of current, and future, neural PDE solvers. We enumerate several evaluation metrics to help rank order the performance of neural PDE solvers. We benchmark the performance of several baseline methods including FNO, CNO, WNO, and DeepONet.

physics.flu-dyn↗