Metraipolis › archive › dc_f55652322f4e

EmbodiedMemory-Bench: Benchmarking Embodied Memory for Long-Horizon Embodied Tasks — arXiv:2609.28236

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.28236
License (as archived)
CC0-1.0
Archived
2026-10-08T12:35:19.563Z
SHA-256 of the archived text
82a03240c8efb8e7fc14af3f48e1d71e748158c3225d2c62fbd2d1cf5b08fe44
Stable URL of this page
https://metraipolis.com/documento/dc_f55652322f4e

How to cite

EmbodiedMemory-Bench: Benchmarking Embodied Memory for Long-Horizon Embodied Tasks — arXiv:2609.28236 — arXiv (en), https://arxiv.org/abs/2609.28236 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f55652322f4e (Metraipolis archive dc_f55652322f4e, sha256 82a03240c8efb8e7fc14af3f48e1d71e748158c3225d2c62fbd2d1cf5b08fe44)
BibTeX
@misc{dcf55652322f4e,
  title = {EmbodiedMemory-Bench: Benchmarking Embodied Memory for Long-Horizon Embodied Tasks — arXiv:2609.28236},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.28236},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f55652322f4e; sha256 82a03240c8efb8e7fc14af3f48e1d71e748158c3225d2c62fbd2d1cf5b08fe44},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.28236. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Long-horizon embodied interaction requires agents to retain and continually update information about the environment as they observe, act, and encounter change. Yet current agents struggle to maintain such memory reliably. Our analysis traces this limitation to four key deficiencies: weak fine-grained visual memory, unreliable dynamic world-state tracking, failing to record world state revealed by interaction outcomes, and limited generalization from prior experience. However, existing benchmarks do not directly assess these memory capabilities during long-horizon embodied interaction. To address this gap, we introduce EmbodiedMemory-Bench (EMem-Bench), comprising 2,554 interactive episodes across four task families. EMem-Bench requires agents to build and update memory from interaction history, then use it to complete a later task by acting in the environment. We further present Embodied-Memorizer (EMem), an external memory system that organizes embodied experience into spatial, event, and scene memories. We also train EMem-8B, an 8B policy that manages and uses these memories. We evaluate a diverse range of open-source and proprietary MLLMs and representative multimodal memory systems. Results show that current models remain weak and uneven across the four challenges. Under matched backbones, EMem achieves the best overall performance among the evaluated memory systems and improves both open-source and proprietary models, while EMem-8B further improves over its backbone. Project page: this https URL

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-24 in cs.CV).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2609.28236 — 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 2026-10-08T12:12:35.794Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-24?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).