Metraipolis › archive › dc_e1445230b2f3

Dual-Latent Memory Routing for Vision-Language Reasoning — arXiv:2609.05539

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.05539
License (as archived)
CC0-1.0
Archived
2026-10-16T15:06:29.947Z
SHA-256 of the archived text
4041b93aa82411a494638fe557ef803967fd241b60bfde6b2d78c0e49026616d
Stable URL of this page
https://metraipolis.com/documento/dc_e1445230b2f3

How to cite

Dual-Latent Memory Routing for Vision-Language Reasoning — arXiv:2609.05539 — arXiv (en), https://arxiv.org/abs/2609.05539 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_e1445230b2f3 (Metraipolis archive dc_e1445230b2f3, sha256 4041b93aa82411a494638fe557ef803967fd241b60bfde6b2d78c0e49026616d)
BibTeX
@misc{dce1445230b2f3,
  title = {Dual-Latent Memory Routing for Vision-Language Reasoning — arXiv:2609.05539},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.05539},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_e1445230b2f3; sha256 4041b93aa82411a494638fe557ef803967fd241b60bfde6b2d78c0e49026616d},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.05539. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Multimodal large language models (MLLMs) have recently made strong progress in vision-language reasoning, yet their performance often degrades as generations grow longer. A key factor is that they frequently lose track of earlier visual evidence and intermediate constraints under a monolithic growing context. Inspired by how humans separately recall what they see and what they infer when solving complex tasks, we propose DLMR, a parameter-efficient mechanism that equips MLLMs with Dual Latent Memories: a visual memory that compresses image evidence and a reasoning memory that tracks intermediate conclusions and constraints. A Router then dynamically decides which memory and how much to reuse during inference, preserving visual grounding while maintaining coherent long-horizon reasoning. DLMR is trained in three stages, from latent memory construction to selective router learning, while keeping the base MLLM frozen, yielding substantial gains on both general and reasoning benchmarks with only a small number of additional trainable parameters. Analyses further show interpretable, state-dependent routing with specialized memory roles and reduced decoding tokens over long generations. Code is available at this https URL .

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-09 in cs.CV).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.05539 — 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 2026-10-16T14:49:56.975Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-09?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).