paper-with-me

홈 › Papers

Multimodal Dataset Distillation via Phased Teacher Models

2026-03-26 · Shengbin Guo, Hang Zhao, Senqiao Yang, Chenyang Jiang, Yuhang Cheng, Xiangru Peng, Rui Shao, Zhuotao Tian arxiv

Multimodal dataset distillation aims to construct compact synthetic datasets that enable efficient compression and knowledge transfer from large-scale image-text data. However, existing approaches often fail to capture the complex, dynamically evolving knowledge embedded in the later training stages of teacher models. This limitation leads to degraded student performance and compromises the quality of the distilled data. To address critical challenges such as pronounced cross-stage performance gaps and unstable teacher trajectories, we propose Phased Teacher Model with Shortcut Trajectory (PTM-ST) -- a novel phased distillation framework. PTM-ST leverages stage-aware teacher modeling and a shortcut-based trajectory construction strategy to accurately fit the teacher's learning dynamics across distinct training phases. This enhances both the stability and expressiveness of the distillation process. Through theoretical analysis and comprehensive experiments, we show that PTM-ST significantly mitigates optimization oscillations and inter-phase knowledge gaps, while also reducing storage overhead. Our method consistently surpasses state-of-the-art baselines on Flickr30k and COCO, achieving up to 13.5% absolute improvement and an average gain of 9.53% on Flickr30k. Code: https://github.com/Previsior/PTM-ST.

📄 PDF Abstract BibTeX arXiv:2603.25388

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

SG-OPD: Sign-Gated On-Policy Distillation via Sign-Consistency Gating and Phased Teacher Sampling

2026-06-08 · Haoran Xu, Hongyu Wang, Yifei Gao, Jiaze Li 외 arxiv

On-policy distillation (OPD) trains a student on its own trajectories with dense per-token supervision from a stronger teacher, and often outperforms off-policy distillation and standard reinforcement learning. However, …

Reinforcement LearningMathematical Reasoning

Phased DMD: Few-step Distribution Matching Distillation via Score Matching within Subintervals

2025-10-31 · Xiangyu Fan, Zesong Qiu, Zhuguanyu Wu, Fanzhou Wang 외 arxiv

Distribution Matching Distillation (DMD) distills score-based generative models into efficient one-step generators, without requiring a one-to-one correspondence with the sampling trajectories of their teachers. Yet, the…

Text-to-Image GenerationVideo Generation

MSD: Saliency-aware Knowledge Distillation for Multimodal Understanding

2021-01-06 · Findings (EMNLP) 2021 11 · Woojeong Jin, Maziar Sanjabi, Shaoliang Nie, Liang Tan 외

To reduce a model size but retain performance, we often rely on knowledge distillation (KD) which transfers knowledge from a large "teacher" model to a smaller "student" model. However, KD on multimodal datasets such as …

Knowledge DistillationMeta-Learning

Uncertainty-Aware Knowledge Distillation for Multimodal Large Language Models

2026-03-22 · Jingchen Sun, Shaobo Han, Deep Patel, Wataru Kohno 외 arxiv

Knowledge distillation establishes a learning paradigm that leverages both data supervision and teacher guidance. However, determining the optimal balance between learning from data and learning from the teacher is chall…

Knowledge Distillation

MST-Distill: Mixture of Specialized Teachers for Cross-Modal Knowledge Distillation

2025-07-09 · Hui Li, Pengfei Yang, Juanyang Chen, Le Dong 외 arxiv

Knowledge distillation as an efficient knowledge transfer technique, has achieved remarkable success in unimodal scenarios. However, in cross-modal settings, conventional distillation methods encounter significant challe…

Knowledge Distillation