@misc{indiciae9dbe3ab41dac,
  title = {Boosting Visual Knowledge-Intensive Training for LVLMs Through Causality-Driven Visual Object Completion},
  author = {Qingguo Hu and Ante Wang and Jia Song and Delai Qiu and Qingsong Liu and Jinsong Su},
  year = {2025},
  url = {https://arxiv.org/abs/2508.04453},
  note = {Source identifier: 2508.04453}
}