Assessing LLMs for Serendipity Discovery in Knowledge Graphs: A Case for Drug Repurposing
Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (19) · 15860-15867
DOI: 10.1609/aaai.v40i19.38618 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Mengying Wang und 8 weiteren Autor:innen.
Abstract (Originalsprache)
Large Language Models (LLMs) have greatly advanced knowledge graph question answering (KGQA), yet existing systems are typically optimized for returning highly relevant but predictable answers. A missing yet desired capacity is to exploit LLMs to suggest surprise and novel ("serendipitious") answers. In this paper, we formally define the serendipity-aware KGQA task and propose the SerenQA framework to evaluate LLMs' ability to uncover unexpected insights in scientific KGQA tasks. SerenQA includes a rigorous serendipity metric based on relevance, novelty, and surprise, along with an expert-annotated benchmark derived from the Clinical Knowledge Graph for drug repurposing. Additionally, it features a structured evaluation pipeline encompassing three subtasks: knowledge retrieval, subgraph reasoning, and serendipity exploration. Our experiments reveal that while state-of-the-art LLMs perform well on retrieval, they still struggle to identify genuinely surprising and valuable discoveries, underscoring a significant room for future research.
Zitieren
APA
Wang, M., Ma, C., Jiao, A., Liang, T., Lu, P., Hegde, S., Yin, Y., Gurkan-Cavusoglu, E., & Wu, Y. (2026). Assessing LLMs for Serendipity Discovery in Knowledge Graphs: A Case for Drug Repurposing. Proceedings of the AAAI Conference on Artificial Intelligence, 40(19), 15860-15867. https://doi.org/10.1609/aaai.v40i19.38618
Vancouver
Wang M, Ma C, Jiao A, Liang T, Lu P, Hegde S, et al. Assessing LLMs for Serendipity Discovery in Knowledge Graphs: A Case for Drug Repurposing. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(19):15860-15867. doi:10.1609/aaai.v40i19.38618
BibTeX
@article{Wang2026,
author = {Wang, Mengying and Ma, Chenhui and Jiao, Ao and Liang, Tuo and Lu, Pengjun and Hegde, Shrinidhi and Yin, Yu and Gurkan-Cavusoglu, Evren and Wu, Yinghui},
title = {Assessing LLMs for Serendipity Discovery in Knowledge Graphs: A Case for Drug Repurposing},
journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
year = {2026},
volume = {40},
number = {19},
pages = {15860--15867},
doi = {10.1609/aaai.v40i19.38618}
}
EUVIMED