Metraipolis › archive › dc_7c61bb3cf346

MAGIC: Margin-gated Contrastive Debiasing with Evidence Restoration for VLM Test-time Robustness Scaling — arXiv:2610.32013

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.32013
License (as archived)
CC0-1.0
Archived
2026-10-26T20:14:02.054Z
SHA-256 of the archived text
712fc10166a826f868f8f3d7aa54f9eafc6fdffca52b1742e54be5b086db9021
Stable URL of this page
https://metraipolis.com/documento/dc_7c61bb3cf346

How to cite

MAGIC: Margin-gated Contrastive Debiasing with Evidence Restoration for VLM Test-time Robustness Scaling — arXiv:2610.32013 — arXiv (en), https://arxiv.org/abs/2610.32013 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7c61bb3cf346 (Metraipolis archive dc_7c61bb3cf346, sha256 712fc10166a826f868f8f3d7aa54f9eafc6fdffca52b1742e54be5b086db9021)
BibTeX
@misc{dc7c61bb3cf346,
  title = {MAGIC: Margin-gated Contrastive Debiasing with Evidence Restoration for VLM Test-time Robustness Scaling — arXiv:2610.32013},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.32013},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7c61bb3cf346; sha256 712fc10166a826f868f8f3d7aa54f9eafc6fdffca52b1742e54be5b086db9021},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.32013. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

This paper studies the problem of test-time robustness scaling for large vision-language models (LVLMs), which aims to mitigate language bias during inference. Existing approaches typically adopt test-time scaling techniques such as contrastive decoding to improve inference performance. Despite the progress, the performance of these approaches is far from satisfactory due to intentional evidence removal during perturbations and redundant inference cost. Towards this end, we propose a novel approach named Margin-gated Contrastive Debiasing with Evidence Restoration (MAGIC). The core of our MAGIC is to evaluate the decision margins of LVLMs to selectively activate textual and visual perturbation branches for efficient debiased inference. In particular, our MAGIC first measures the prediction margins of crucial tokens, and applies a gating function to identify low-confidence responses that are susceptible to language biases. For selected samples, we integrate the token logits from multiple textual and image perturbation branches with the original ones for debiased decision. More importantly, we selectively restore the most informative corrupted views with sharp prediction distributions for ambiguous samples after debiasing, enabling evidence recovery for discriminative decision-making. We further utilize a calibration dataset to actively identify and prune ineffective branches during contrastive debiasing for cost reduction. Extensive experiments on benchmark datasets validate the effectiveness of the proposed MAGIC in comparison with extensive baselines.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.32013 — 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 2026-10-26T19:49:03.459Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).