paper-with-me

Papers

VAD: Attributing Visual Evidence for Target Reconstruction in Multimodal On-Policy Distillation

2026-07-30 · Kangning Zhang, Yixing Li, Shuai Shao, Qingyao Li, Zhengxi Lu, Zhiyuan Yao, Jianghao Lin, Wenxiang Jiao, Yuan Lu, Weiwen Liu, Weinan Zhang, Yong Yu arxiv

Multimodal on-policy distillation (OPD) transfers fine-grained visual knowledge by supervising student-generated trajectories with a privileged-view teacher. Yet its next-token corrections are source-mixed, combining visual signals with linguistic priors and teacher-specific effects. The key challenge is to estimate which corrections are supported by visual evidence, not merely where or how strongly to distill. We introduce Visual Attribution Distillation (VAD), a counterfactual target-reconstruction algorithm that estimates the visually attributable part of a teacher correction. At each student-generated prefix, VAD evaluates the same fixed teacher with the relevant evidence present and removed. The corresponding change in centered log-probabilities defines ut, a signed proxy for the visual evidence direction that estimates how revealing the evidence supports or refutes candidate tokens. VAD projects the original correction onto this proxy to obtain an intervention-aligned component and a proxy-unexplained residual, then reconstructs a student-anchored target from the former. During training, this reconstructed target supplies the primary supervision signal, while the privileged teacher contributes a weak regularizer. Across six fine-grained visual benchmarks at 4B and 9B scales, VAD outperforms direct privileged-view distillation and visual-advantage weighting. Token- level and controlled-target analyses show that the proxy-aligned component is enriched in task-relevant visual corrections and yields stronger target shifts, especially when evidence refutes a mistaken answer. These results support counterfactual target reconstruction as an effective alternative to source-mixed supervision.

📄 PDF Abstract BibTeX arXiv:2607.28590

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

PAL-Bench: Evidence-Grounded Profile Reconstruction from Longitudinal Personal Albums

2026-06-15 · Qiwei Yan, Zhiqiang Yuan, Zexi Jia, Nanxing Hu 외 arxiv

Longitudinal personal albums are weak-schema multimodal databases: noisy perceptual records whose key facts require joins across faces, text, timestamps, locations, and repeated events. Existing visual, video, document, …

Structured PredictionEntity Resolution

Paint What You See: Benchmarking Dexterous Visual Tool Use in Multimodal Agents

2026-08-26 · Shudong Liu, Dongyang Chen, Enci Zhang, Jinwei Liang 외 arxiv

Evaluation is shifting from static QA toward agentic settings where models act through external tools. We identify a critical yet underexplored capability within this space - dexterous visual tool use: fine-grained, clos…

V-Retrver: Evidence-Driven Agentic Reasoning for Universal Multimodal Retrieval

2026-02-05 · Dongyang Chen, Chaoyang Wang, Dezhao Su, Xi Xiao 외 arxiv

Multimodal Large Language Models (MLLMs) have recently been applied to universal multimodal retrieval, where Chain-of-Thought (CoT) reasoning improves candidate reranking. However, existing approaches remain largely lang…

Reinforcement Learning

Multimodal Speech Enhancement Using Burst Propagation

2022-09-07 · Mohsin Raza, Leandro A. Passos, Ahmed Khubaib, Ahsan Adeel

This paper proposes the MBURST, a novel multimodal solution for audio-visual speech enhancements that consider the most recent neurological discoveries regarding pyramidal cells of the prefrontal cortex and other brain r…

ManagementSpeech Enhancement

CRAFT: Critic-Refined Adaptive Key-Frame Targeting for Multimodal Video Question Answering

2026-05-18 · Mahesh Bhosale, Abdul Wasi, Vishvesh Trivedi, Pengyu Yan 외 arxiv

Grounded multi-video question answering over real-world news events requires systems to surface query-relevant evidence across heterogeneous video archives while attributing every claim to its supporting source. We intro…

Video Question Answering