Metraipolis › archive › dc_bf9f69c5ea0b

SemReward-VL: Semantic Reward-Guided Video-Language Adaptation for Developmental Behavior Assessment — arXiv:2609.33082

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.33082
License (as archived)
CC0-1.0
Archived
2026-10-07T21:00:57.428Z
SHA-256 of the archived text
ed8b198f0dbcbb849cc0723bd7655cf62ae382204833a489eb61f9011828da22
Stable URL of this page
https://metraipolis.com/documento/dc_bf9f69c5ea0b

How to cite

SemReward-VL: Semantic Reward-Guided Video-Language Adaptation for Developmental Behavior Assessment — arXiv:2609.33082 — arXiv (en), https://arxiv.org/abs/2609.33082 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_bf9f69c5ea0b (Metraipolis archive dc_bf9f69c5ea0b, sha256 ed8b198f0dbcbb849cc0723bd7655cf62ae382204833a489eb61f9011828da22)
BibTeX
@misc{dcbf9f69c5ea0b,
  title = {SemReward-VL: Semantic Reward-Guided Video-Language Adaptation for Developmental Behavior Assessment — arXiv:2609.33082},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.33082},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_bf9f69c5ea0b; sha256 ed8b198f0dbcbb849cc0723bd7655cf62ae382204833a489eb61f9011828da22},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.33082. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Developmental screening videos show how children perform specific behaviors, but clinical records usually contain outcomes rather than descriptions of what happened. We present SemReward-VL, which learns to describe item-specific behavior from these outcomes. A vision-language model generates a description, and a frozen language model scores its agreement with the clinical outcome, relevance to the item, abstention on unrelated video-item pairs, and clarity. Group relative policy optimization (GRPO) updates LoRA adapters using this semantic reward. On 13,379 videos covering 41 items, the method improves aggregate accuracy and the number of items with recall above 0.5. Errors remain in temporal direction, duration, and age-specific interpretations of behavior.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.AI).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.33082 — 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 2026-10-07T20:29:02.535Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).