Mnemosyne: Accelerating Multi-Hop Question Answering via Cache Hit Order Fitting
Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (36) · 30557-30565
DOI: 10.1609/aaai.v40i36.40310 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Haizhou Du und 4 weiteren Autor:innen.
Abstract (Originalsprache)
Multi-Hop Question Answering (MHQA) requires step-by-step reasoning across multiple pieces of information to answer complex questions. The cache-aided Retrieval-Augmented Generation (RAG) can accelerate the process of external knowledge retrieval at each reasoning step for MHQA. However, existing methods focus on the internal structure and ignore the misalignment between the queries’ arrival order and cache hit order. To tackle this, we propose Mnemosyne, a cache hit order fitting method designed to accelerate the RAG progress for MHQA. Specifically, our cache-aware order fitting strategy adjusts the order of queries arrival via graph reordering to better align with the cache hit order, thereby reducing the likelihood of failed or unproductive retrieval attempts. The multi-granularity caching storage mechanism is designed to loosen the strict hit condition to multiple similar semantic matching modes, facilitating that relevant documents can still be retrieved. Experiments conducted on four multi-hop QA datasets demonstrate that Mnemosyne effectively reduces retrieval latency while enhancing task answer F1 score, achieving a superior trade-off between efficiency and effectiveness.
Zitieren
APA
Du, H., Li, J., Li, D., Huang, L., & Wang, L. (2026). Mnemosyne: Accelerating Multi-Hop Question Answering via Cache Hit Order Fitting. Proceedings of the AAAI Conference on Artificial Intelligence, 40(36), 30557-30565. https://doi.org/10.1609/aaai.v40i36.40310
Vancouver
Du H, Li J, Li D, Huang L, Wang L. Mnemosyne: Accelerating Multi-Hop Question Answering via Cache Hit Order Fitting. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(36):30557-30565. doi:10.1609/aaai.v40i36.40310
BibTeX
@article{Du2026,
author = {Du, Haizhou and Li, Jiujiu and Li, Dongyang and Huang, Luobin and Wang, Lisheng},
title = {Mnemosyne: Accelerating Multi-Hop Question Answering via Cache Hit Order Fitting},
journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
year = {2026},
volume = {40},
number = {36},
pages = {30557--30565},
doi = {10.1609/aaai.v40i36.40310}
}
EUVIMED