# Astrolabe: Steering Forward-Process Reinforcement Learning for Distilled Autoregressive Video Models

arXiv: https://arxiv.org/abs/2603.17051 · HTML: https://mlsanity.com/paper/2603.17051/

## Key facts

- Cited by 15 papers in the map (data through 14 Oct 2026).
- Placed in RL for robot locomotion and manipulation, one of 15 directions in Deep reinforcement learning methods.
- First posted on arXiv 17 Mar 2026.

## Placed in

- Deep reinforcement learning methods / [RL for robot locomotion and manipulation](https://mlsanity.com/deeprl/rl-for-robot-locomotion-and-manipulation/)

## Built on by (most cited first, 15 of 15)

- [Matrix-Game 3.0: Real-Time and Streaming Interactive World Model with Long-Horizon Memory](https://mlsanity.com/paper/2604.08995/) (2026)
- [DreamX-World 1.0: A General-Purpose Interactive World Model](https://mlsanity.com/paper/2606.16993/) (2026)
- [SolarWM: Open Data and Scalable Training for Long-Horizon Video World Models](https://mlsanity.com/paper/2609.02886/) (2026)
- [KVPO: ODE-Native GRPO for Autoregressive Video Alignment via KV Semantic Exploration](https://mlsanity.com/paper/2605.14278/) (2026)
- [RAVEN: Real-time Autoregressive Video Extrapolation with Consistency-model GRPO](https://mlsanity.com/paper/2605.15190/) (2026)
- [A Systematic Post-Train Framework for Video Generation](https://mlsanity.com/paper/2604.25427/) (2026)
- [TempAct: Advancing Temporal Plausibility in Autoregressive Video Generation via Planner-Executor RL](https://mlsanity.com/paper/2606.28016/) (2026)
- [From Pixels to States: Rethinking Interactive World Models as Game Engines](https://mlsanity.com/paper/2607.14076/) (2026)
- [PhyMotion: Structured 3D Motion Reward for Physics-Grounded Human Video Generation](https://mlsanity.com/paper/2605.14269/) (2026)
- [On-Policy Adversarial Flow Distillation for Autoregressive Video Generation](https://mlsanity.com/paper/2605.26105/) (2026)
- [Focused Forcing: Content-Aware Per-Frame KV Selection for Efficient Autoregressive Video Diffusion](https://mlsanity.com/paper/2605.18346/) (2026)
- [Mask Forcing: Improving Autoregressive Video Diffusion Distillation via Dual-Noise Masking Rollout](https://mlsanity.com/paper/2609.09123/) (2026)
- [FlexComposer: Unified Video Compositing from Images to Dynamic Footage with Flexible Trajectory Control](https://mlsanity.com/paper/2607.29627/) (2026)
- [Stream4D: 4D-Consistency for Streaming Autoregressive Diffusion Video Models](https://mlsanity.com/paper/2608.19556/) (2026)
- [OneFixer: High-Quality and Consistent One-Step Autoregressive 3DGS Refinement for Driving Scenes](https://mlsanity.com/paper/2609.32175/) (2026)

## Abstract (arXiv)

Distilled autoregressive (AR) video models enable efficient streaming generation but frequently misalign with human visual preferences. Existing reinforcement learning (RL) frameworks are not naturally suited to these architectures, typically requiring either expensive re-distillation or solver-coupled reverse-process optimization that introduces considerable memory and computational overhead. We present Astrolabe, an efficient online RL framework tailored for distilled AR models. To overcome existing bottlenecks, we introduce a forward-process RL formulation based on negative-aware fine-tuning. By contrasting positive and negative samples directly at inference endpoints, this approach establishes an implicit policy improvement direction without requiring reverse-process unrolling. To scale this alignment to long videos, we propose a streaming training scheme that generates sequences progressively via a rolling KV-cache, applying RL updates exclusively to local clip windows while conditioning on prior context to ensure long-range coherence. Finally, to mitigate reward hacking, we integrate a multi-reward objective stabilized by uncertainty-aware selective regularization and dynamic reference updates. Extensive experiments demonstrate that our method consistently enhances generation quality across multiple distilled AR video models, serving as a robust and scalable alignment solution.