LectureRechercheRadarCadre d'investissement
Connexion / Inscription
← Retour à la lecture
ActualitésSujet · VisTA视觉弃权模型Matériel1 min de lectureRédaction DisconfirmAI

Multimodal Models Learn to Refuse Impossible Edits

New research introduces visual abstention; the VisTA method enables models to refuse 93% of infeasible drawing requests while maintaining high editing accuracy.

2026-10-07 12:00 GMT+8

Unified multimodal models (UMMs) can now identify impossible drawing instructions and actively refuse them, rather than generating erroneous images.

Previously, even the strongest editing models refused only 0.4% of logically infeasible modification requests, often hallucinating non-existent objects or silently altering the prompt. Explicitly prompting models to report infeasibility increased refusals but reduced editing accuracy.

The VisTA training method proposed by Chufan Shi et al. pairs feasible and infeasible examples, forcing the model to judge feasibility before generating output. The resulting VisTA-BAGEL model refuses 93.0% of infeasible requests without reminders, while completing 74.3% of feasible edits—outperforming all eight evaluated mainstream models.

These results rely on the authors' self-created Draw-or-Decline benchmark and have not yet been independently reproduced.

Sources

  • arxiv.org↗
Sujet

VisTA视觉弃权模型

南加州大学与加州大学圣地亚哥分校团队提出VisTA(Visual Transformation and Abstention)训练方法,旨在解决统一多模态模型(UMMs)在图像编辑中无法识别“不可行指令”的问题。该研究正式定义了“视觉弃权”…

2 étapes

Connexion / Inscription
LectureRechercheRadarCadre d'investissement
1 pièces
Mis à jour 2026-10-08 13:14

Comment on en est arrivé là

  1. 2026-10-06论文《Visual Abstention in Unified Multimodal Models》提交至arXiv,提出DoD基准和VisTA方法。arxiv.org ↗
  2. 2026-10-08新闻发布指出VisTA使模型能识别并拒绝93%的不可行绘图请求,同时保持高编辑准确率。arxiv.org ↗

Ce qui reste non résolu

  • VisTA方法是否适用于其他类型的多模态生成任务(如视频编辑)?
  • 在实际部署中,如何平衡高拒绝率带来的用户体验影响与安全收益?
Lire le sujet complet →
Continuer la recherche →