Metraipolis › archive › dc_6d54686ec070
VISTA: Internalizing Collective Visual Experience via On-Policy Distillation for Active Multimodal Agents — arXiv:2609.38086
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.38086
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T18:41:42.558Z
- SHA-256 of the archived text
9eafdaef30b1bad72c24368e69e0b8fb4542ec36684dc07820f802e3020f2a59- Stable URL of this page
https://metraipolis.com/documento/dc_6d54686ec070
How to cite
VISTA: Internalizing Collective Visual Experience via On-Policy Distillation for Active Multimodal Agents — arXiv:2609.38086 — arXiv (en), https://arxiv.org/abs/2609.38086 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_6d54686ec070 (Metraipolis archive dc_6d54686ec070, sha256 9eafdaef30b1bad72c24368e69e0b8fb4542ec36684dc07820f802e3020f2a59)
BibTeX
@misc{dc6d54686ec070,
title = {VISTA: Internalizing Collective Visual Experience via On-Policy Distillation for Active Multimodal Agents — arXiv:2609.38086},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.38086},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_6d54686ec070; sha256 9eafdaef30b1bad72c24368e69e0b8fb4542ec36684dc07820f802e3020f2a59},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.38086. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Active multimodal agents use visual tools to acquire task-relevant evidence while reasoning. Although reinforcement learning samples multiple interaction trajectories per input, outcome-based objectives primarily use the group to estimate scalar advantages, leaving complementary visual discoveries underused. We introduce VISTA, which internalizes collective visual experience through on-policy distillation by turning observations from same-input rollouts into shared supervision. Collective visual experience distillation (CVED) organizes these observations with their interaction context and aligns them with individual decisions, while heterogeneity-aware policy improvement (HAPI) reinforces successful trajectories and provides experience-guided distillation for unsuccessful attempts. An experience-conditioned teacher evaluates the student's sampled response prefixes, allowing discoveries from one trajectory to guide learning in another without replacing the student's original history or generating new target trajectories. The trained agent retains its visual tools and acts using its own interaction history. VISTA achieves the strongest average performance among the evaluated active multimodal agents of comparable size and consistently outperforms same-backbone training baselines across fine-grained perception and general reasoning tasks, demonstrating the value of collective experience for active multimodal learning. Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-30 in cs.CV). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV). Pagina del lavoro: https://arxiv.org/abs/2609.38086 — 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 2026-10-07T18:04:24.938Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-30?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).