Mitigating Error Accumulation in Knowledge Editing for Multi-Hop Question Answering
Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (37) · 30862-30870
DOI: 10.1609/aaai.v40i37.40344 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Jiaxin Guo und 4 weiteren Autor:innen.
Abstract (Originalsprache)
Knowledge editing (KE) has emerged as an effective approach for updating factual information in large language models (LLMs) without the need for full retraining. Most of the existing methods for addressing the "ripple effect" in KE adopt a chain-structured reasoning process, making them vulnerable to error accumulation from early incorrect steps. Moreover, their conflict detection mechanisms are often susceptible to the LLM's inherent confirmation bias, further undermining the reliability of the editing process. To overcome these challenges, we propose Tree of Editing (ToE), a tree-structured, retrieval-enhanced knowledge editing framework designed to support robust reasoning under factual updates. ToE expands reasoning paths using a breadth-first strategy combined with score-guided beam search, enabling diverse and error-tolerant inference. Besides, we introduce an observer to objectively update knowledge, avoiding the bias caused by LLMs' over-confidence. Experimental results on two benchmarks, namely MQuAKE-CF (targeting ripple-aware editing) and DUNE (free-form editing), demonstrate that ToE framework significantly outperforms existing methods.
Zitieren
APA
Guo, J., Sun, H., Zhang, W., Fan, X., & Zhang, Y. (2026). Mitigating Error Accumulation in Knowledge Editing for Multi-Hop Question Answering. Proceedings of the AAAI Conference on Artificial Intelligence, 40(37), 30862-30870. https://doi.org/10.1609/aaai.v40i37.40344
Vancouver
Guo J, Sun H, Zhang W, Fan X, Zhang Y. Mitigating Error Accumulation in Knowledge Editing for Multi-Hop Question Answering. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(37):30862-30870. doi:10.1609/aaai.v40i37.40344
BibTeX
@article{Guo2026,
author = {Guo, Jiaxin and Sun, Hao and Zhang, Wenhao and Fan, Xuanbo and Zhang, Yan},
title = {Mitigating Error Accumulation in Knowledge Editing for Multi-Hop Question Answering},
journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
year = {2026},
volume = {40},
number = {37},
pages = {30862--30870},
doi = {10.1609/aaai.v40i37.40344}
}
EUVIMED