A Progressive Training Strategy for Vision-Language Models to Counteract Spatio-Temporal Hallucinations in Embodied Reasoning

Xiaoda Yang, Shuai Yang, Can Wang, Jingyang Xue, Menglan Tang, Checheng Yu, Xunzhe Zhou, Sashuai Zhou, Tao Jin, Lixin Yang, Xiangyu Yue, Zhou ZhaoarXiv 2026 · DOI

BibTeX
@misc{yang2026progressive,
  author = {Xiaoda Yang and Shuai Yang and Can Wang and Jingyang Xue and Menglan Tang and Checheng Yu and Xunzhe Zhou and Sashuai Zhou and Tao Jin and Lixin Yang and Xiangyu Yue and Zhou Zhao},
  title = {{A Progressive Training Strategy for Vision-Language Models to Counteract Spatio-Temporal Hallucinations in Embodied Reasoning}},
  year = {2026},
  doi = {10.48550/arxiv.2604.10506},
}

Loading related content...

Figure