Constant-Memory Differentiable Light Tracing

Linas Beresna; Eugene Fiume

Home/Archive/arXiv

arXiv · 2609.32920

Constant-Memory Differentiable Light Tracing

Linas Beresna·Eugene Fiume

View original source ↗Read source document ↗Export metadata ↓

Abstract

Reverse-mode differentiation of Monte Carlo light transport naively requires storing a computation graph whose size grows with path length, making it impractical for deep or high-sample simulations. Path Replay Backpropagation (PRB) eliminates this memory complexity for viewpoint path tracing by reconstructing paths from their random seeds and performing a constant-memory backward replay. However, this formulation relies on a one-path-one-pixel property that does not hold for light tracing: a single light path can contribute to many pixels, and the corresponding per-vertex adjoints cannot be recovered during replay from constant-size state. We demonstrate that this structural difference prevents a direct extension of PRB to light tracing in constant memory, and that buffered two-pass variants retain linear memory scaling with path length. We then introduce two constant-memory reverse-mode formulations for differentiable light tracing. The first, Reservoir Light Replay Backpropagation (ResLRB), compresses the set of per-path sensor connections into a single stochastically selected representative using weighted reservoir sampling, preserving a two-pass structure at the cost of increased gradient variance. The second, LRB-3-pass, introduces an additional traversal that accumulates downstream adjoint contributions prior to backpropagation, retaining all connections and matching naive reverse-mode AD in expectation. Both methods support detached and attached formulations, including differentiation of sensor splat positions induced by geometric perturbations. We validate correctness against naive AD, characterize memory and runtime trade-offs, and demonstrate inverse rendering of specular caustics driven entirely by light tracing.

Explore related subjects

cs.GR

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Linas Beresna, Eugene Fiume. 2026-09-26. Constant-Memory Differentiable Light Tracing. https://arxiv.org/abs/2609.32920

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Squeeze3D: Extreme Neural Compression with Latent Space Bridging

We propose Squeeze3D, a novel framework that leverages implicit prior knowledge learnt by existing pre-trained encoders and decoders to compress 3D data at extremely high compression ratios. Our approach bridges the latent spaces between a pre-trained encoder and a pretrained decoder model through trainable mapping networks. Any 3D asset represented as a mesh, point cloud, or radiance field is first encoded by the pre-trained encoder and then transformed (i.e. compressed) into a highly compact latent code by a mapping network. This latent code can effectively be used as an extremely compressed representation of the mesh, point cloud, or radiance field. A mapping network transforms the compressed latent code into the latent space of a powerful generative model; the decoder of this generative model then recreates the original 3D asset (i.e. decompression). Squeeze3D is trained entirely on generated synthetic data and does not require any 3D datasets. The Squeeze3D architecture can be flexibly used with existing pre-trained 3D encoders and existing generative models. It can flexibly support different formats, including meshes, point clouds, and radiance fields. Our experiments demonstrate that Squeeze3D achieves compression ratios of up to 2187$\times$ for textured meshes, 58.5$\times$ for point clouds, and more than 650$\times$ for radiance fields while maintaining visual quality comparable to many existing methods. Squeeze3D only incurs a small compression and decompression latency since it does not involve training object-specific networks to compress an object.

cs.GR↗

arXiv2026

ViscoReg: Neural Signed Distance Functions via Viscosity Solutions

Implicit Neural Representations (INRs) that learn Signed Distance Functions (SDFs) from point cloud data represent the state-of-the-art for geometrically accurate 3D scene reconstruction. However, training these Neural SDFs involves enforcing the Eikonal equation, an ill-posed equation that also leads to unstable gradient flows. Numerical Eikonal solvers have relied on viscosity approaches for regularization and stability. Motivated by this well-established theory, we introduce ViscoReg, a novel regularizer for Neural SDF methods, and theoretically prove that it stabilizes training. Empirically, ViscoReg outperforms state-of-the-art approaches such as SIREN, DiGS, StEik, and HotSpot across most metrics on ShapeNet, the Surface Reconstruction Benchmark, 3D scene reconstruction and reconstruction from real scans. We also establish novel generalization error estimates for Neural SDFs in terms of the training error, using the theory of viscosity solutions. Our empirical and theoretical results provide confidence in the general applicability of our method.

cs.GR↗

arXiv2026

Printing the Underdetermined: Materializing Multi-solutionness in Figurative Paintings

Figurative paintings are often approached as if they depict a single recoverable 3D scene: viewers infer depth and occlusion, and reconstruction pipelines attempt to converge to one stable model. We instead foreground multi-solutionness, the non-uniqueness of 3D configurations compatible with a single painted image, and propose a workflow that keeps this non-uniqueness visible and material. Multi-solutionness arises from two sources: unobserved content, where backsides and occluded volumes admit multiple plausible completions, and observed cues, where perspective, shading, and occlusion still underconstrain geometry. When additional views are synthesized by a video generative model without explicit 3D constraints, small frame-level drifts become inevitable rather than exceptional. Our pipeline samples multiple camera-orbit multi-view video sequences from one painting, reconstructs each sequence with 3D Gaussian Splatting into a point-based Gaussian scene representation where density halos and ghosting expose unresolved degrees of freedom, and fabricates these representations as physical artifacts using DreamPrinting. By treating multiple compatible interpretations as explicit outputs rather than residual error, we provide a computational framework for spatial readings of figurative painting that can be inspected, compared, and discussed in both digital and physical form.

cs.GR↗