Newest papers

Looking Ahead: Modeling Future Evolution in Autoregressive Video Generation

Ruopeng Gao, Yukun Huang, Jiarui Hu, Shuai Wang, Wenbo Hu, Ying Shan, Limin Wang · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.29989 · cited by 0 works (OpenAlex) · open-access PDF

German

Autoregressive Video-Difusionsmodelle erzeugen Videos von Stück zu Stück, aber die Erzeugung plausibler Fortsetzungen garantiert nicht, dass sie erfassen, wie Szenen sich entwickeln. Unter Lehrerzwingung kann eine starke zeitliche Redundanz die lokale visuelle Kontinuität als Abkürzung für die aktuelle Chunk-Denoisierung dienen, wodurch der Druck auf die visuelle Evolutionsmodelle verringert wird. Wir führen Future Chunk Prediction (FCP) ein, ein Hilfsmodell, das gemeinsam mit Lehrer-Zwangs-Denoisierung optimiert wird, die zeitlich getrennte Stücke vorhersagt, ohne Zugang zu intervenierenden visuellen Kontexten zu haben. Aufsicht direkt aus Trainingsvideos ziehen, FCP erfordert keinen zusätzlichen vor

Machine translation (Meta M2M100); the English original follows.

Abstract

Autoregressive video diffusion models generate videos chunk by chunk, yet producing plausible continuations does not guarantee that they capture how scenes evolve. Under teacher forcing, strong temporal redundancy may allow local visual continuity to serve as a shortcut for current-chunk denoising, reducing the pressure to model visual evolution. We introduce Future Chunk Prediction (FCP), an auxiliary objective jointly optimized with teacher-forced denoising that predicts temporally separated chunks without access to the intervening visual context. Drawing supervision directly from training videos, FCP requires no additional pretrained visual encoder and leaves autoregressive inference unchanged. Our method achieves superior physical plausibility over standard teacher-forced autoregressive training with a lower overall training compute budget, highlighting the effectiveness and efficiency of future-chunk supervision for modeling visual evolution. Intriguingly, our visualizations suggest that FCP fosters a more structured representation space, which may be valuable for building autoregressive world models.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Autoregressive model, Computer science, Redundancy (engineering), Inference, Artificial intelligence, Encoder

Is this paper real? This page was built from OpenAlex and arXiv records, not generated. Software can verify any reference in one call: GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.29989, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txt

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange