EUVIMEDEUVIMED entdecken
Crossref

On the Superimposed Noise Accumulation Problem in Sequential Knowledge Editing of Large Language Models

Ding Cao, Yuchen Cai, Yuqing Huang, Xuesong He, Rongxi Guo, Guiquan Liu, Guangzhong Sun

Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (36) · 30139-30147

DOI: 10.1609/aaai.v40i36.40263 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Ding Cao und 6 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Sequential knowledge editing techniques aim to continuously update knowledge in large language models at low cost, preventing models from generating outdated or incorrect information. However, existing sequential editing methods suffer from a significant decline in editing success rates after long-term editing. Through theoretical analysis and experiments, our findings reveal that as the number of edits increases, the model's output increasingly deviates from the desired target, leading to a drop in editing success rates. We refer to this issue as the superimposed noise accumulation problem. Our further analysis demonstrates that the problem is related to the erroneous activation of irrelevant knowledge and conflicts between activated knowledge. Based on this analysis, a method named DeltaEdit is proposed that reduces conflicts between knowledge through dynamic orthogonal constraint strategies. Experiments show that DeltaEdit significantly reduces superimposed noise, achieving a 16.8% improvement in editing performance over the strongest baseline.

Zitieren

APA

Cao, D., Cai, Y., Huang, Y., He, X., Guo, R., Liu, G., & Sun, G. (2026). On the Superimposed Noise Accumulation Problem in Sequential Knowledge Editing of Large Language Models. Proceedings of the AAAI Conference on Artificial Intelligence, 40(36), 30139-30147. https://doi.org/10.1609/aaai.v40i36.40263

Vancouver

Cao D, Cai Y, Huang Y, He X, Guo R, Liu G, et al. On the Superimposed Noise Accumulation Problem in Sequential Knowledge Editing of Large Language Models. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(36):30139-30147. doi:10.1609/aaai.v40i36.40263

BibTeX
@article{Cao2026,
  author = {Cao, Ding and Cai, Yuchen and Huang, Yuqing and He, Xuesong and Guo, Rongxi and Liu, Guiquan and Sun, Guangzhong},
  title = {On the Superimposed Noise Accumulation Problem in Sequential Knowledge Editing of Large Language Models},
  journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
  year = {2026},
  volume = {40},
  number = {36},
  pages = {30139--30147},
  doi = {10.1609/aaai.v40i36.40263}
}
Kostenlos starten →