Metraipolis › archive › dc_7fa7283cb2d8

Who Drives the Probability Game of VLMs? A Temporal Causal Drive Evaluation Framework — arXiv:2609.02000

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.02000
License (as archived): CC0-1.0
Archived: 2026-10-17T05:49:49.011Z
SHA-256 of the archived text: 495e707d76ca059ec0de8c54fdcfa239b6571efce9a753d2ee079cc0756c8860
Stable URL of this page: https://metraipolis.com/documento/dc_7fa7283cb2d8

How to cite

Who Drives the Probability Game of VLMs? A Temporal Causal Drive Evaluation Framework — arXiv:2609.02000 — arXiv (en), https://arxiv.org/abs/2609.02000 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7fa7283cb2d8 (Metraipolis archive dc_7fa7283cb2d8, sha256 495e707d76ca059ec0de8c54fdcfa239b6571efce9a753d2ee079cc0756c8860)

BibTeX

@misc{dc7fa7283cb2d8,
  title = {Who Drives the Probability Game of VLMs? A Temporal Causal Drive Evaluation Framework — arXiv:2609.02000},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.02000},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7fa7283cb2d8; sha256 495e707d76ca059ec0de8c54fdcfa239b6571efce9a753d2ee079cc0756c8860},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.02000. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Vision-language models (VLMs) are increasingly evaluated on complex image and video understanding tasks, yet conventional metrics primarily assess final-answer quality and reveal little about how different information sources shape the generation process. We propose a causal and temporal evaluation framework that traces the evolving roles of visual input, question text, and generated prefixes during autoregressive decoding. Grounded in a Structural Causal Model, we use interventions and backdoor adjustment to derive three step-indexed causal-drive metrics---Visual Causal Drive (VCD), Question Causal Drive (QCD), and Prefix Causal Drive (PCD)---for characterizing source-specific generation patterns without requiring reference answers. Experiments on Qwen3-VL-8B-Instruct across MAVIS, LLaVA-Video-178K, and MiraData, together with cross-model validation on InternVL2-8B, reveal a consistent transition from stronger early question and visual guidance toward increasing reliance on generated prefixes. Randomized-intervention validation shows that QCD and PCD reduce recovery error over observational PMI baselines by 34.8\% and 47.1\%, respectively. On VLMBias, the prefix--visual imbalance score achieves 0.767 AUROC and 0.873 AUPRC for distinguishing prior-driven from visually grounded generations. These results show that causal-drive trajectories provide complementary source-level diagnostics for multimodal generation.

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-03 in cs.CV).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2609.02000 — 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 2026-10-17T05:25:58.049Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-03?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).