AI Paper Atlas

Research map / 3D vision / 3D generation

DLWM: Dual Latent World Models enable Holistic Gaussian-centric Pre-training in Autonomous Driving

Yiyao Zhu, Ying Xue, Haiming Zhang, Guangfeng Jiang, Wending Zhou, Xu Yan, Jiantao Gao, Yingjie Cai, Bingbing Liu, Zhen Li, Shaojie Shen

CVPR 2026 · 3D vision › 3D generation · approach: 3D perception for autonomous driving · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add two-stage pre-trainingDLWM · 3D occupancy perceptionmIoU: 20.83 -> 21.85Improved
add two-stage pre-trainingDLWM · 4D occupancy forecastingmIoU: 15.09 -> 17.77Improved
add two-stage pre-trainingDLWM · motion planningL2: 0.55 -> 0.46Improved
add dense pseudo-depth supervision (ablation)DLWM · perception ablationmIoU: 18.40 -> 18.74Improved
add semantic supervision (ablation)DLWM · perception ablationmIoU: 18.74 -> 18.99Improved
add Gaussian-flow-guided latent world model (stage 2) (ablation)DLWM · perception ablationmIoU: 18.99 -> 19.30Improved
use latent BEV instead of raw Gaussian features for planning (ablation)DLWM · motion planningL2: 0.55 -> 0.52Improved
add future BEV forecasting to planning (ablation)DLWM · motion planningL2: 0.52 -> 0.50Improved
add stage-1 reconstruction pre-training to planning (ablation)DLWM · motion planningL2: 0.50 -> 0.46Improved
use dual decoupled world models instead of a unified one (ablation)DLWM · planningL2: 0.58 -> 0.46Improved
raise Gaussian count from 25600 to 51200 (ablation)DLWM · 3D occupancy perceptionIoU: 30.56 -> 29.79Mixed
predict 1 future frame instead of 0 (ablation)DLWM · 3D occupancy perceptionmIoU: 18.83 -> 19.30Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 6 of 6 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy