An Advanced Gradient Leakage Attack Against Duplicate Labels via Model Outputs Reconstruction

Kunlan Xiang, Haomiao Yang, Meng Hao 0001, Zikang Ding, Hongwei Li 0001, Qingchuan Zhao, Tianwei Zhang 0004. An Advanced Gradient Leakage Attack Against Duplicate Labels via Model Outputs Reconstruction. IEEE Trans. Dependable Sec. Comput., 23(3):5875-5888, May - June 2026. [doi]

@article{XiangYHDLZZ26,
  title = {An Advanced Gradient Leakage Attack Against Duplicate Labels via Model Outputs Reconstruction},
  author = {Kunlan Xiang and Haomiao Yang and Meng Hao 0001 and Zikang Ding and Hongwei Li 0001 and Qingchuan Zhao and Tianwei Zhang 0004},
  year = {2026},
  month = {May - June},
  doi = {10.1109/TDSC.2026.3659196},
  url = {https://doi.org/10.1109/TDSC.2026.3659196},
  researchr = {https://researchr.org/publication/XiangYHDLZZ26},
  cites = {0},
  citedby = {0},
  journal = {IEEE Trans. Dependable Sec. Comput.},
  volume = {23},
  number = {3},
  pages = {5875-5888},
}