Metraipolis › archive › dc_a09006e02af0

EM^2Mem: Event-Centric Multimodal Memory for Large Language Models — arXiv:2609.00551

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.00551
License (as archived)
CC0-1.0
Archived
2026-10-17T06:33:29.255Z
SHA-256 of the archived text
719be4af0772510d18a1c944162b5dff8e143ad467b86902fb1b371ff3c44ae8
Stable URL of this page
https://metraipolis.com/documento/dc_a09006e02af0

How to cite

EM^2Mem: Event-Centric Multimodal Memory for Large Language Models — arXiv:2609.00551 — arXiv (en), https://arxiv.org/abs/2609.00551 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_a09006e02af0 (Metraipolis archive dc_a09006e02af0, sha256 719be4af0772510d18a1c944162b5dff8e143ad467b86902fb1b371ff3c44ae8)
BibTeX
@misc{dca09006e02af0,
  title = {EM^2Mem: Event-Centric Multimodal Memory for Large Language Models — arXiv:2609.00551},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.00551},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_a09006e02af0; sha256 719be4af0772510d18a1c944162b5dff8e143ad467b86902fb1b371ff3c44ae8},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.00551. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Multimodal memory offers a scalable interface for long-video question answering, but existing methods often retrieve captions, frames, transcripts, summaries, or graph facts as isolated fragments. Although searchable, such fragments are not generation-ready: language models must reconstruct cross-modal and temporal alignments at inference time, when context is limited and attribution is difficult. We propose EM^2Mem, an event-centric multimodal memory framework that binds heterogeneous evidence to event anchors during memory construction. Each event-indexed memory cell aligns multimodal records, temporal context, graph-linked relations, semantic facts, and provenance, enabling compact evidence readout over grounded multimodal events rather than modality-specific fragments. Across three long-video QA benchmarks, EM^2Mem improves average accuracy over the strongest memory baseline by 2.0, 2.4, and 3.7 points, improves strict event-level Top-5 evidence recall by 7.0 points, and reduces per-query latency by 4.67 times and total inference tokens by 63.66% (The code will be integrated into this https URL ).

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-02 in cs.AI).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Multimedia (cs.MM).
Pagina del lavoro: https://arxiv.org/abs/2609.00551 — 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 2026-10-17T06:15:12.798Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).