Backdoor Watermarking for Continuous Prompt Learning Models in Industrial Systems

Abstract

Large Language Models (LLMs) are becoming key enablers in adaptive and autonomous systems, particularly under the paradigm of Industry 5.0, where human-centric design and generative Artificial Intelligence (AI) technologies are increasingly deployed. However, the widespread of LLMs raises serious intellectual property concerns, especially in few-shot learning scenarios where model customization is achieved through continuous prompt tuning. Traditional watermarking methods fail to protect

Categories

Framework mappings

Suggested from the entry's categories.

@misc{chen2026backdoor,
  title = {{Backdoor Watermarking for Continuous Prompt Learning Models in Industrial Systems}},
  author = {Kongyang Chen and Chuwen Pang and Xiaolin Wang and Tiancai Liang and Jiaxing Shen},
  year = {2026},
  month = jun,
  doi = {10.1145/3820766},
  url = {https://doi.org/10.1145/3820766}
}