Delta Forcing: Trust Region Steering for Interactive Autoregressive Video Generation

TL;DR

Delta Forcing introduces a trust region-based framework for autoregressive video generation, balancing responsiveness to new events with long-term temporal coherence by constraining teacher supervision.

Contribution

It proposes a novel trust region approach inspired by policy optimization to improve stability and reactivity in interactive video generation models.

Findings

Significantly improves temporal consistency in generated videos.

Maintains responsiveness to dynamic event conditions.

Outperforms existing methods in experimental evaluations.

Abstract

Interactive real-time autoregressive video generation is essential for applications such as content creation and world modeling, where visual content must adapt to dynamically evolving event conditions. A fundamental challenge lies in balancing reactivity and stability: models must respond promptly to new events while maintaining temporal coherence over long horizons. Existing approaches distill bidirectional models into autoregressive generators and further adapt them via streaming long tuning, yet often exhibit persistent drift after condition changes. We identify the cause as conditional bias, where the teacher may provide condition-aligned but trajectory-agnostic guidance, biasing generation toward locally valid yet globally inconsistent modes. Inspired by Trust Region Policy Optimization, we propose Delta Forcing, a simple yet effective framework that constrains unreliable teacher supervision within an adaptive trust region. Specifically, Delta Forcing estimates transition consistency from the latent delta between teacher and generator trajectories, and uses it to balance teacher supervision with a monotonic continuity objective. This suppress unreliable teacher-induced shifts while preserving responsiveness to new events. Extensive experiments demonstrate that Delta Forcing significantly improves consistency while maintaining event reactivity.