Metraipolis › archive › dc_49ee50ac3be0

MemoryCPT: An End-to-End Agent Memory Framework for Cost-Performance Trade-off — arXiv:2608.04843

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.04843
License (as archived)
CC0-1.0
Archived
2026-10-31T11:47:39.597Z
SHA-256 of the archived text
d90a1d2cacb565fb4b4df9fcedecd40553e4ccaf049ff8514879c3e2a4d084e1
Stable URL of this page
https://metraipolis.com/documento/dc_49ee50ac3be0

How to cite

MemoryCPT: An End-to-End Agent Memory Framework for Cost-Performance Trade-off — arXiv:2608.04843 — arXiv (en), https://arxiv.org/abs/2608.04843 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_49ee50ac3be0 (Metraipolis archive dc_49ee50ac3be0, sha256 d90a1d2cacb565fb4b4df9fcedecd40553e4ccaf049ff8514879c3e2a4d084e1)
BibTeX
@misc{dc49ee50ac3be0,
  title = {MemoryCPT: An End-to-End Agent Memory Framework for Cost-Performance Trade-off — arXiv:2608.04843},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.04843},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_49ee50ac3be0; sha256 d90a1d2cacb565fb4b4df9fcedecd40553e4ccaf049ff8514879c3e2a4d084e1},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.04843. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Long-horizon LLM agents require memory systems that recover useful evidence from large interaction histories without passing excessive context to downstream models. Existing memory pipelines often rely on hand-crafted heuristics and repeated LLM calls, which can introduce redundant context and high inference cost. We propose MemoryCPT, an end-to-end trainable agent memory pipeline that spans offline memory construction and online query-conditioned context generation. MemoryCPT consists of two stages: Query-agnostic Distillation (QAD), which distills a modular memory-construction pipeline into a compact model using explicit reasoning traces; and Query-aware Retrieval and Summarization (QAR), which combines reciprocal rank fusion (RRF) with a LoRA-based summarizer trained via Group Relative Policy Optimization (GRPO) under a cost-aware reward. We further introduce Quality per Cost (QPC) to quantify answer quality per unit inference cost. Experiments on LoCoMo and LongMemEval show that MemoryCPT improves the cost-performance trade-off over the evaluated baselines, while ablation and sensitivity analyses characterize the contributions of its components and the effects of key design choices.

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-08-06 in cs.IR).
Categorie arXiv: Information Retrieval (cs.IR).
Pagina del lavoro: https://arxiv.org/abs/2608.04843 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-31T11:37:02.910Z dalla pagina https://arxiv.org/catchup/cs.IR/2026-08-06?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).