Metraipolis › archive › dc_c4fec08caa07
MUSE: Multi-Level Memory with Selective Runtime Injection for Scientific Coding Agents — arXiv:2610.38332
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.38332
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-31T06:12:42.241Z
- SHA-256 of the archived text
3ad0b3001074be241cca22a007cd6928d35af938753c0ff6a62d157a3ee436d4- Stable URL of this page
https://metraipolis.com/documento/dc_c4fec08caa07
How to cite
MUSE: Multi-Level Memory with Selective Runtime Injection for Scientific Coding Agents — arXiv:2610.38332 — arXiv (en), https://arxiv.org/abs/2610.38332 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_c4fec08caa07 (Metraipolis archive dc_c4fec08caa07, sha256 3ad0b3001074be241cca22a007cd6928d35af938753c0ff6a62d157a3ee436d4)
BibTeX
@misc{dcc4fec08caa07,
title = {MUSE: Multi-Level Memory with Selective Runtime Injection for Scientific Coding Agents — arXiv:2610.38332},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.38332},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_c4fec08caa07; sha256 3ad0b3001074be241cca22a007cd6928d35af938753c0ff6a62d157a3ee436d4},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.38332. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Historical experience can guide scientific coding at multiple levels, but task-level retrieval may miss implementation lessons that become relevant during execution. We introduce MUSE, a memory management framework combining multi-level memory with selective runtime injection for scientific coding agents. MUSE builds global, approach-level, and implementation-level memory from historical trajectories. At initialization, the agent receives global memory and problem-matched approach-level memory. During execution, current interactions are used to retrieve implementation-level memory. An LLM gate selects relevant experience and generates specific review advice, or skips injection. Across four benchmarks spanning numerical problem solving, kernel optimization, and scientific software repair, MUSE outperforms retrieval and memory baselines on all evaluation metrics. Ablations support the multi-level design and gated injection, while transfer experiments show that memory remains useful across kernel benchmarks. These results show that multi-level memory and selective runtime injection enable effective reuse of experience in scientific coding. Code is available at this https URL . Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2610.38332 — 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 2026-10-31T06:07:08.648Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).