Metraipolis › archive › dc_f1698dd63af3
Video-HopChain: Multi-Hop Questions and Confidence-Gated Exploration for Video Reasoning Models — arXiv:2609.25773
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.25773
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-08T13:46:38.880Z
- SHA-256 of the archived text
fb6ef1a01d52e6e290849925e27178e8280fe3fae61c73bc4688c939cd551178- Stable URL of this page
https://metraipolis.com/documento/dc_f1698dd63af3
How to cite
Video-HopChain: Multi-Hop Questions and Confidence-Gated Exploration for Video Reasoning Models — arXiv:2609.25773 — arXiv (en), https://arxiv.org/abs/2609.25773 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f1698dd63af3 (Metraipolis archive dc_f1698dd63af3, sha256 fb6ef1a01d52e6e290849925e27178e8280fe3fae61c73bc4688c939cd551178)
BibTeX
@misc{dcf1698dd63af3,
title = {Video-HopChain: Multi-Hop Questions and Confidence-Gated Exploration for Video Reasoning Models — arXiv:2609.25773},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.25773},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f1698dd63af3; sha256 fb6ef1a01d52e6e290849925e27178e8280fe3fae61c73bc4688c939cd551178},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.25773. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
HopChain has shown on still images that multi-hop data synthesis improves vision-language reasoning, because long chain-of-thought reasoning exposes errors that compound across steps, while most data used for reinforcement learning with verifiable rewards (RLVR) rarely demands a chain of visual evidence, so these weaknesses are likely to stay unexposed. We observe the same problem in video, where this framework has not yet been explored. We therefore build Video-HopChain, a dataset of 22,550 multi-hop video questions over 13,378 videos, together with a held-out benchmark of 1,000 questions. Each question chains three to six yes/no questions about moments in one video, and each yields one of two integers depending on its answer. The final answer is the sum of these integers, so an exact match on that sum gives the verifiable reward that RLVR needs. We first train Qwen3-VL-8B with GRPO on a standard video dataset, and a second stage on Video-HopChain then raises the mean over eight video understanding and reasoning benchmarks from 55.4 to 57.9 and improves every one of them. Training on such a dataset, however, exposes a known limitation of GRPO: its learning signal comes from the reward variance within a group, so hard questions whose rollouts are all incorrect and easy questions whose rollouts are all correct both leave the group with no gradient. To recover these groups at the same compute budget, we introduce Confidence-Gated Exploration (CGE). With 8 rollouts per question, CGE samples the first 4 as usual. If these 4 are either all correct or all incorrect, it samples the last 4 with the policy's most confident token masked inside the reasoning span, and removes the masked positions from the loss while all 8 rollouts enter the advantage. With CGE, the mean rises further to 59.3. We release the dataset, the checkpoint, and the data generation and training code. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-23 in cs.AI). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.25773 — 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 2026-10-08T13:18:38.259Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-23?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).