Metraipolis › archive › dc_4074fed88201
Grounding Memory Summarization in Utility Intent — arXiv:2609.33417
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.33417
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T21:53:01.368Z
- SHA-256 of the archived text
e9bd38b2c8f7c7bd9f5957fdd83589ad601fec11776c7ea7f2bade44035e2720- Stable URL of this page
https://metraipolis.com/documento/dc_4074fed88201
How to cite
Grounding Memory Summarization in Utility Intent — arXiv:2609.33417 — arXiv (en), https://arxiv.org/abs/2609.33417 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_4074fed88201 (Metraipolis archive dc_4074fed88201, sha256 e9bd38b2c8f7c7bd9f5957fdd83589ad601fec11776c7ea7f2bade44035e2720)
BibTeX
@misc{dc4074fed88201,
title = {Grounding Memory Summarization in Utility Intent — arXiv:2609.33417},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.33417},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_4074fed88201; sha256 e9bd38b2c8f7c7bd9f5957fdd83589ad601fec11776c7ea7f2bade44035e2720},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.33417. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Existing summarizers for memory systems are typically optimized for human-facing criteria such as faithfulness, which misaligns with their true objective: preserving the evidence needed to support future queries. We show that conditioning summarization on query-answer pairs substantially improves answer quality, and that this utility-aware behavior is transferable across queries. Motivated by these findings, we propose MemSuit, a self-distillation framework in which a teacher summarizer, conditioned on observed query-answer pairs, produces utility-aware memory entries that a student learns to reproduce from the raw conversation alone. To prevent collateral erasure where conditioning on a single query-answer pair discards evidence relevant to other plausible queries, the teacher decomposes each block into multiple self-contained entries that preserve distinct query-relevant facets as independently retrievable units. To align the retriever with the compact, fact-dense style of teacher entries, we further fine-tune the embedding model with a contrastive objective supervised by teacher entries. Across a diverse suite of conversational query types, MemSuit consistently outperforms state-of-the-art baselines, confirming the value of grounding memory in downstream utility. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.CL). Categorie arXiv: Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2609.33417 — 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 2026-10-07T21:27:08.518Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).