AURORA: Automated Training Framework of Universal Process Reward Models via Ensemble Prompting and Reverse Verification

Xiaoyu Tan, Tianchu Yao, Chao Qu, Bin Li, Minghao Yang, Dakuan Lu, Haozhe Wang, Yinghui Xu, Xihe Qiu

BibTeX
@inproceedings{tan2026aurora,
  author = {Xiaoyu Tan and Tianchu Yao and Chao Qu and Bin Li and Minghao Yang and Dakuan Lu and Haozhe Wang and Yinghui Xu and Xihe Qiu},
  title = {{AURORA: Automated Training Framework of Universal Process Reward Models via Ensemble Prompting and Reverse Verification}},
  booktitle = {ACM SIGKDD Conference on Knowledge Discovery and Data Mining},
  year = {2026},
  doi = {10.1145/3770854.3780168},
}
Figure