EmboAlign: Aligning Video Generation with Compositional Constraints for Zero-Shot Manipulation

TL;DR

EmboAlign is a novel framework that leverages vision-language models to impose compositional constraints on video generative models, enabling more accurate and safe zero-shot robotic manipulation without task-specific training.

Contribution

The paper introduces a data-free method that aligns video generative models with constraints from vision-language models at inference time for improved robotic manipulation.

Findings

Improves success rate by 43.3 percentage points on real robot tasks.

Effectively filters and refines VGM outputs using constraint-guided selection.

Enhances zero-shot manipulation accuracy without additional training data.

Abstract

Video generative models (VGMs) pretrained on large-scale internet data can produce temporally coherent rollout videos that capture rich object dynamics, offering a compelling foundation for zero-shot robotic manipulation. However, VGMs often produce physically implausible rollouts, and converting their pixel-space motion into robot actions through geometric retargeting further introduces cumulative errors from imperfect depth estimation and keypoint tracking. To address these challenges, we present \method{}, a data-free framework that aligns VGM outputs with compositional constraints generated by vision-language models (VLMs) at inference time. The key insight is that VLMs offer a capability complementary to VGMs: structured spatial reasoning that can identify the physical constraints critical to the success and safety of manipulation execution. Given a language instruction, \method{} uses a VLM to automatically extract a set of compositional constraints capturing task-specific requirements, which are then applied at two stages: (1) constraint-guided rollout selection, which scores and filters a batch of VGM rollouts to retain the most physically plausible candidate, and (2) constraint-based trajectory optimization, which uses the selected rollout as initialization and refines the robot trajectory under the same constraint set to correct retargeting errors. We evaluate \method{} on six real-robot manipulation tasks requiring precise, constraint-sensitive execution, improving the overall success rate by 43.3\% points over the strongest baseline without any task-specific training data.