Safety
VARI: Robust Text-to-Video Generation via Group-Structured Intent Alignment
arXiv:2610.18189v1 Announce Type: new Abstract: Modern text-to-video (T2V) models generate visually compelling videos, yet meaning-preserving changes in prompt wording can substantially alter semantic
arXiv:2610.18189v1 Announce Type: new Abstract: Modern text-to-video (T2V) models generate visually compelling videos, yet meaning-preserving changes in prompt wording can substantially alter semantic realization and overall quality. Existing prompt rewriting modifies the inference input, whereas conventional post-training constructs supervision from one prompt formulation at a time; neither explicitly optimizes reliability across equivalent forms. We introduce VARI (Verified Alignment over Representations of Intent), a group-structured framework that treats automatically verified formulations of the same visual intent as a shared supervision unit. VARI builds matched form-seed rollout pools, cross-scores each video against all forms, and applies a wording-independent structured intent audit to identify valid realizations and hard semantic failures. Group-Consistent Distillation (GCD) aligns every verified form with interchangeable valid realizations, while Form-Contrastive Flow-DPO (FC-DPO) learns group-level preferences from audited positive-negative pairs, prioritizing same-seed, different-form contrasts. On HunyuanVideo-1.5, VARI attains the highest T2V-CompBench, VBench, and StoryEval scores among the evaluated methods. On prompt-form extensions of T2V-CompBench and VBench, it improves weakest-form alignment over the strongest prompt-level baselines by 2.45 and 3.55 points and reduces cross-form drop by 30.5% and 32.2%, respectively, while preserving visual quality. Human evaluation, natural-caption controls, prompt-enhancement experiments, diversity analyses, and transfer to LTX-2.3 further support the robustness and transferability of these gains, while VARI requires no additional inference-time module.
Related
- Adversarial Gradient Forcing for Robust Few-Step Autoregressive Video Generation
- CAPE-T2V: Captioner-Anchored Prompt Enhancement toward Two-Sided Conditioning Alignment in Text-to-Video Generation
- RAPO++: Cross-Stage Prompt Optimization for Text-to-Video Generation via Data Alignment and Test-Time Scaling
Source: arXiv cs.CV | 2026-10-14