Metraipolis › archive › dc_d0046ec6aa1a

VaME: Exploring Variational Latent Reasoning for Multimodal Embeddings — arXiv:2609.33402

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.33402
License (as archived)
CC0-1.0
Archived
2026-10-07T23:51:35.766Z
SHA-256 of the archived text
805fe7f6096ca72309b87bda3eb5842f1ba555c1737790278099a7d0d5698eff
Stable URL of this page
https://metraipolis.com/documento/dc_d0046ec6aa1a

How to cite

VaME: Exploring Variational Latent Reasoning for Multimodal Embeddings — arXiv:2609.33402 — arXiv (en), https://arxiv.org/abs/2609.33402 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d0046ec6aa1a (Metraipolis archive dc_d0046ec6aa1a, sha256 805fe7f6096ca72309b87bda3eb5842f1ba555c1737790278099a7d0d5698eff)
BibTeX
@misc{dcd0046ec6aa1a,
  title = {VaME: Exploring Variational Latent Reasoning for Multimodal Embeddings — arXiv:2609.33402},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.33402},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d0046ec6aa1a; sha256 805fe7f6096ca72309b87bda3eb5842f1ba555c1737790278099a7d0d5698eff},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.33402. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Universal multimodal retrieval requires compact embeddings that preserve task-relevant semantic information across diverse modalities. Prior works have incorporated latent reasoning into multimodal embedding learning to refine this information before embedding extraction. However, most existing approaches remain confined to deterministic latent paths, without exploring alternative trajectories to discover better embeddings. Thus, we propose VaME (Variational Multimodal Embeddings), a framework that models latent reasoning as a learnable distribution over trajectories. Specifically, we first introduce Variational Latent Reasoning (VLR) to enable autoregressive exploration in latent space, guided by answer reconstruction through a lightweight decoder. Meanwhile, we augment the original embedding-token readout with a latent-fused embedding to facilitate exploration during subsequent reinforcement learning. Finally, we optimize latent reasoning over stochastic variational trajectories through reinforcement learning, using Semantic Decoding Reward (SDR) to favor semantically meaningful trajectories with interpretable decoded outcomes. On the 78-task MMEB-V2 benchmark, spanning image, video, and visual-document retrieval, VaME outperforms most explicit CoT-based models and all latent-reasoning baselines. VaME also demonstrates robust performance on reasoning-intensive benchmarks such as MRMR, with substantial gains after reinforcement learning. Importantly, VaME achieves these gains with at least a 4.25x inference speedup over the deterministic latent autoregressive baselines. The code will be made publicly available.

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-29 in cs.CV).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2609.33402 — 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 2026-10-07T23:29:27.202Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).