arXiv:2510.24718v2 Announce Type: replace-cross 
Abstract: Autoregressive video diffusion models are capable of long rollouts that are stable and consistent with history, but they are unable to guide the current generation with conditioning from the future. In camera-guided video generation with a predefined camera trajectory, this limitation leads to collisions with the generated scene, after which autoregression quickly collapses. To address this, we propose Generative View Stitching (GVS), which samples the entire sequence in parallel such that the generated scene is faithful to every part of the predefined camera trajectory. Our main contribution is a sampling algorithm that extends prior work on diffusion stitching for robot planning to video generation. While such stitching methods usually require a specially trained model, GVS is compatible with any off-the-shelf video model trained with Diffusion Forcing, a prevalent sequence diffusion framework that we show already provides the affordances necessary for stitching. We then introduce Omni Guidance, a technique that enhances the temporal consistency in stitching by conditioning on both the past and future, and that enables our proposed loop-closing mechanism for delivering long-range coherence. Overall, GVS achieves camera-guided video generation that is stable, collision-free, frame-to-frame consistent, and closes loops for a variety of predefined camera paths, including Oscar Reutersv\"ard's Impossible Staircase. Results are best viewed as videos at https://andrewsonga.github.io/gvs.

تم اقتراح نهج جديد يسمى Generative View Stitching (GVS) لتحسين توليد الفيديو من خلال معالجة قيود النماذج الذاتية الانحدار. هذه النماذج، على الرغم من فعاليتها في إنشاء مقاطع فيديو مستقرة، تواجه صعوبات في التكييف المستقبلي، مما يؤدي غالبًا إلى عدم الاتساق في المشاهد المولدة. يهدف GVS إلى تحسين ذلك من خلال دمج تقنيات موجهة بالكاميرا، مما يسمح بإنتاج مقاطع فيديو أكثر اتساقًا وواقعية. هذه الخطوة مهمة لأنها قد تحدث ثورة في كيفية إنشاء المحتوى المرئي والتفاعل معه، مما يجعله أكثر غمرًا وتوافقًا مع توقعات المستخدمين.

Se ha propuesto un nuevo enfoque llamado Generative View Stitching (GVS) para mejorar la generación de video al abordar las limitaciones de los modelos autorregresivos. Estos modelos, aunque efectivos en la creación de videos estables, tienen dificultades con el condicionamiento futuro, lo que a menudo conduce a inconsistencias en las escenas generadas. GVS busca mejorar esto integrando técnicas guiadas por cámara, lo que permite salidas de video más coherentes y realistas. Este avance es significativo ya que podría revolucionar la forma en que creamos e interactuamos con el contenido de video, haciéndolo más inmersivo y alineado con las expectativas del usuario.

Une nouvelle approche appelée Generative View Stitching (GVS) a été proposée pour améliorer la génération vidéo en abordant les limitations des modèles autorégressifs. Ces modèles, bien qu'efficaces pour créer des vidéos stables, ont du mal avec le conditionnement futur, ce qui entraîne souvent des incohérences dans les scènes générées. GVS vise à améliorer cela en intégrant des techniques guidées par la caméra, permettant des sorties vidéo plus cohérentes et réalistes. Cette avancée est significative car elle pourrait révolutionner notre façon de créer et d'interagir avec le contenu vidéo, le rendant plus immersif et conforme aux attentes des utilisateurs.

A new approach called Generative View Stitching (GVS) has been proposed to enhance video generation by addressing the limitations of autoregressive models. These models, while effective in creating stable video rollouts, struggle with future conditioning, often leading to inconsistencies in generated scenes. GVS aims to improve this by integrating camera-guided techniques, allowing for more coherent and realistic video outputs. This advancement is significant as it could revolutionize how we create and interact with video content, making it more immersive and aligned with user expectations.

Generative View Stitching

arXiv:2601.07894v1 Announce Type: new 
Abstract: Masked diffusion models (MDMs), which leverage bidirectional attention and a denoising process, are narrowing the performance gap with autoregressive models (ARMs). However, their internal attention mechanisms remain under-explored. This paper investigates the attention behaviors in MDMs, revealing the phenomenon of Attention Floating. Unlike ARMs, where attention converges to a fixed sink, MDMs exhibit dynamic, dispersed attention anchors that shift across denoising steps and layers. Further analysis reveals its Shallow Structure-Aware, Deep Content-Focused attention mechanism: shallow layers utilize floating tokens to build a global structural framework, while deeper layers allocate more capability toward capturing semantic content. Empirically, this distinctive attention pattern provides a mechanistic explanation for the strong in-context learning capabilities of MDMs, allowing them to double the performance compared to ARMs in knowledge-intensive tasks. All codes and datasets are available at https://github.com/NEUIR/Attention-Floating.

كشفت دراسة حديثة عن آلية الانتباه العائم في نماذج الانتشار المقنعة (MDMs)، موضحة سلوكيات الانتباه الفريدة التي تختلف عن النماذج الذاتية التكرار التقليدية (ARMs). تُظهر هذه الأبحاث أن نماذج MDMs تستخدم نقاط انتباه ديناميكية تتحرك عبر الطبقات وخطوات إزالة الضوضاء، مما يساهم في تحسين أدائها في المهام التي تتطلب التعلم في السياق.

Un estudio reciente ha revelado el mecanismo de Atención Flotante en los Modelos de Difusión Enmascarados (MDMs), destacando sus comportamientos de atención únicos que difieren de los modelos autorregresivos tradicionales (ARMs). Esta investigación muestra que los MDMs utilizan anclas de atención dinámicas que se desplazan a través de capas y pasos de desruido, contribuyendo a su rendimiento mejorado en tareas que requieren aprendizaje en contexto.

Une étude récente a révélé le mécanisme d'Attention Floating dans les Modèles de Diffusion Masqués (MDMs), mettant en lumière leurs comportements d'attention uniques qui diffèrent des modèles autoregressifs traditionnels (ARMs). Cette recherche montre que les MDMs utilisent des ancres d'attention dynamiques qui se déplacent à travers les couches et les étapes de débruitage, contribuant à leur performance améliorée dans les tâches nécessitant un apprentissage contextuel.

A recent study has unveiled the Attention Floating mechanism in Masked Diffusion Models (MDMs), highlighting their unique attention behaviors that differ from traditional autoregressive models (ARMs). This research reveals that MDMs utilize dynamic attention anchors that shift across layers and denoising steps, contributing to their enhanced performance in tasks requiring in-context learning.

Generative View Stitching

Was this article worth reading? Share it

LucidQuery AI

Videotok

AiReelGenerator.com

Postugc

Videolulu

Cococlip.AI

Ready to build your own newsroom?