arXiv:2606.12987v1 Announce Type: new 
Abstract: Action-conditioned world models let an autonomous vehicle predict future camera scenes from its own planned controls, enabling planning and simulation without real-world rollouts, but at compact, trainable scale the futures are ambiguous and the field's standard distortion metrics actively mislead: they reward a blurry regression mean over a realistic prediction. We confront this with a compact latent world model that, given the present front-camera latent and a sequence of ego-actions, predicts future scene latents a frozen decoder renders to $256 \times 256$ frames up to 8 seconds ahead, evaluated on 150 held-out nuScenes scenes. We first benchmark where to predict: across six frozen encoders spanning four representation families, V-JEPA2 with temporal context reduces steering RMSE by 40% over the best single-frame encoder. We then train a latent Diffusion Transformer (DiT) and, through a controlled diagnosis, identify the four ingredients it needs: spatial tokens, the $x_0$ objective, residual anchoring, and sampling matched to target uncertainty. In a Stable-Diffusion-VAE encode-predict-decode pipeline we expose the central tension: distortion metrics (cosine similarity, SSIM) favor the blurry mean, masking that the diffusion model is far closer to the real frame distribution. Inception-based FID and KID reveal a clean perception-distortion frontier: diffusion attains KID 0.078 versus 0.375 for regression ($4.8\times$ better), and a deployable train-derived calibration makes this practical without test-time ground truth. The model is genuinely action-controllable (steering drives scene displacement, Spearman $\rho = 0.81$, vs $-0.18$ for regression). We trace limited single-pass motion to a shared-present anchor and engineer a compact 1.7M-parameter "jump" model that recovers full ground-truth motion magnitude ($1.02\times$ GT), where single-pass models capture less than half.

تقدم دراسة جديدة نموذج Diffusion Transformer World-Action، الذي يمكّن المركبات الذاتية القيادة من توقع مشاهد الكاميرا المستقبلية بناءً على التحكم المخطط، مما يعزز قدرات التخطيط والمحاكاة دون الحاجة إلى تنفيذات في العالم الحقيقي. يستخدم هذا النموذج نموذج عالم كامن مضغوط للتنبؤ بكامنة المشهد المستقبلية، محققًا تحسينات كبيرة في دقة التنبؤ.

Un nuevo estudio presenta el modelo Diffusion Transformer World-Action, que permite a los vehículos autónomos predecir futuras escenas de cámara basándose en los controles planificados, mejorando así las capacidades de planificación y simulación sin despliegues en el mundo real. Este modelo utiliza un modelo de mundo latente compacto para prever los latentes de escena futuros, logrando mejoras significativas en la precisión de las predicciones.

Une nouvelle étude présente le modèle Diffusion Transformer World-Action, qui permet aux véhicules autonomes de prédire les futures scènes de caméra en fonction des commandes prévues, améliorant ainsi les capacités de planification et de simulation sans déploiements réels. Ce modèle utilise un modèle de monde latent compact pour prévoir les latents de scène futurs, réalisant des améliorations significatives dans la précision des prédictions.

A new study introduces the Diffusion Transformer World-Action Model, which enables autonomous vehicles to predict future camera scenes based on planned controls, enhancing planning and simulation capabilities without real-world rollouts. This model utilizes a compact latent world model to forecast future scene latents, achieving significant improvements in prediction accuracy.