cd /news/computer-vision/split-and-scale-latent-3d-representa… · home › topics › computer-vision › article
[ARTICLE · art-171410] src=machinelearning.apple.com ↗ pub= topic=computer-vision verified=true sentiment=↑ positive

Split-and-Scale Latent 3D Representations

Researchers Dorian Chan, Muhammed Kocabas, Xiaoming Zhao, Oncel Tuzel and Jen-Hao Rick Chang published a paper on split-and-scale, an inference-time method that expands the latent capacity of 3D representations without raising training cost, by applying a single pretrained tokenizer to arbitrarily small sub-regions of a scene. The authors report that extensive ablations show the approach achieves reconstruction quality competitive with much larger models that exceed single-GPU memory limits, and that an autoregressive 3D generator trained on the representation performs competitively with state-of-the-art baselines.

recommendation

Scale latent capacity at inference, not training: reuse a single pretrained tokenizer on smaller scene sub-regions to gain detail without exceeding single-GPU memory.

more in playbook: Inference and self-hosting
read2 min views14 publishedNov 16, 2026
Split-and-Scale Latent 3D Representations
Image: Apple ML Research

content type paperpublished November 2026 Split-and-Scale Latent 3D Representations

AuthorsDorian Chan, Muhammed Kocabas, Xiaoming Zhao, Oncel Tuzel, Jen-Hao Rick Chang

Latent 3D representations have significantly expanded the capabilities of modern shape modeling, enabling compact encoding of 3D scenes and supporting powerful 3D generative models. Despite this progress, current representations struggle to capture the fine details of complex objects, fundamentally limited by the capacity of their latents. However, scaling latent capacity at training time is computationally expensive and often infeasible under memory limits, and naively increasing the number of latents at inference fails to make effective use of the added capacity. In this paper, we address both challenges with split-and-scale, a method that effectively expands latent capacity at inference while keeping training cost tractable. Our key observation is that by representing 3D scenes as continuous functions of 3D coordinates, a single pretrained tokenizer can be applied at inference to arbitrarily small sub-regions of a scene, allocating the same latent capacity to a smaller spatial extent and thus capturing far higher detail. We validate this insight by conducting extensive ablations to study the trade-offs between compute, memory, latent capacity, and reconstruction quality, demonstrating our inference-time scaling method achieves quality competitive with much larger models that exceed single-GPU memory limits. We further demonstrate that split-and-scale supports high-quality generative modeling, training an autoregressive 3D generator that performs competitively with state-of-the-art baselines.

High-fidelity image-to-3D generation requires a 3D representation that captures both geometry and appearance. To support relighting and integration into standard rendering pipelines, the representation should include physically based rendering (PBR) modalities such as albedo, metallic-roughness, and surface normals. We propose Luce, a 3D representation that unifies geometry and PBR materials within a voxelized multimodal Gaussian cloud, using…

GAUDI: A Neural Architect for Immersive 3D Scene Generation

September 27, 2022research area Computer Vision, research area Methods and Algorithmsconference NeurIPS

We introduce GAUDI, a generative model capable of capturing the distribution of complex and realistic 3D scenes that can be rendered immersively from a moving camera. We tackle this challenging problem with a scalable yet powerful approach, where we first optimize a latent representation that disentangles radiance fields and camera poses. This latent representation is then used to learn a generative model that enables both unconditional and…

── more in #computer-vision 4 stories · sorted by recency
── more on @dorian chan 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/split-and-scale-late…] indexed:0 read:2min 2026-11-16 · —