MESA learns a query-adaptive subset of five memory structures for long-horizon agents, beating AMA-Agent by 8.5 points on AMA-Bench while using 41% fewer evidence tokens than reading all structures.
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers) , pages=
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
MESA:Task-Adaptive Multi-Structure Evidence Selection for Long-Horizon Agent Memory
MESA learns a query-adaptive subset of five memory structures for long-horizon agents, beating AMA-Agent by 8.5 points on AMA-Bench while using 41% fewer evidence tokens than reading all structures.