DCARL: A Divide-and-Conquer Framework for Autoregressive Long-Trajectory Video Generation

2026-03-25

AI Summary

Existing video generation methods face significant challenges in synthesizing long-duration videos, including the limited scalability of diffusion models and the visual drift and poor controllability of autoregressive models. This work proposes DCARL, a novel framework that introduces a divide-and-conquer strategy to autoregressive long-form video generation. DCARL employs a keyframe generator to establish globally consistent structural anchors and integrates overlapping-segment autoregressive interpolation to preserve local temporal coherence while leveraging the high-fidelity generation capabilities of video diffusion models. Trained on large-scale long-trajectory video datasets, DCARL substantially outperforms current state-of-the-art methods, achieving superior performance across multiple metrics—including FID, FVD, ATE, and ARE—and enables stable synthesis of high-fidelity videos up to 32 seconds in length.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Natural Language Processing: Generation

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Long-trajectory video generation is a crucial yet challenging task for world modeling primarily due to the limited scalability of existing video diffusion models (VDMs). Autoregressive models, while offering infinite rollout, suffer from visual drift and poor controllability. To address these issues, we propose DCARL, a novel divide-and-conquer, autoregressive framework that effectively combines the structural stability of the divide-and-conquer scheme with the high-fidelity generation of VDMs. Our approach first employs a dedicated Keyframe Generator trained without temporal compression to establish long-range, globally consistent structural anchors. Subsequently, an Interpolation Generator synthesizes the dense frames in an autoregressive manner with overlapping segments, utilizing the keyframes for global context and a single clean preceding frame for local coherence. Trained on a large-scale internet long trajectory video dataset, our method achieves superior performance in both visual quality (lower FID and FVD) and camera adherence (lower ATE and ARE) compared to state-of-the-art autoregressive and divide-and-conquer baselines, demonstrating stable and high-fidelity generation for long trajectory videos up to 32 seconds in length.

Problem

• long-trajectory video generation
• autoregressive models
• visual drift
• controllability
• video diffusion models

Innovation

• Divide-and-Conquer
• Autoregressive Video Generation
• Keyframe-Guided Interpolation
• Long-Trajectory Video Modeling
• Video Diffusion Models