Paper Detail
Huanxuan Liao, Zhongtao Jiang, Yupu Hao, Yuqiao Tan, Shizhu He, Jun Zhao, Kun Xu, Kang Liu
Multimodal Large Language Models (MLLMs) achieve stronger visual understanding by scaling input fidelity, yet the resulting visual token growth makes jointly sustaining high spatial resolution and long temporal context prohibitive. We argue that the bottleneck lies not in how post-encoding representations are compressed but in the volume of pixels the encoder receives, and address it with ResAdapt, an Input-side adaptation framework that learns how much visual budget each frame should receive before encoding. ResAdapt couples a lightweight Allocator with an unchanged MLLM backbone, so the backbone retains its native visual-token interface while receiving an operator-transformed input. We formulate allocation as a contextual bandit and train the Allocator with Cost-Aware Policy Optimization (CAPO), which converts sparse rollout feedback into a stable accuracy-cost learning signal. Across budget-controlled video QA, temporal grounding, and image reasoning tasks, ResAdapt improves low-budget operating points and often lies on or near the efficiency-accuracy frontier, with the clearest gains on reasoning-intensive benchmarks under aggressive compression. Notably, ResAdapt supports up to 16x more frames at the same visual budget while delivering over 15% performance gain. Code is available at https://github.com/Xnhyacinth/ResAdapt.
No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.
No ranking explanation is available yet.
No tags.
@article{liao2026resadapt,
title = {ResAdapt: Adaptive Resolution for Efficient Multimodal Reasoning},
author = {Huanxuan Liao and Zhongtao Jiang and Yupu Hao and Yuqiao Tan and Shizhu He and Jun Zhao and Kun Xu and Kang Liu},
year = {2026},
abstract = {Multimodal Large Language Models (MLLMs) achieve stronger visual understanding by scaling input fidelity, yet the resulting visual token growth makes jointly sustaining high spatial resolution and long temporal context prohibitive. We argue that the bottleneck lies not in how post-encoding representations are compressed but in the volume of pixels the encoder receives, and address it with ResAdapt, an Input-side adaptation framework that learns how much visual budget each frame should receive be},
url = {https://arxiv.org/abs/2603.28610},
keywords = {cs.CV, cs.AI, cs.CL, Multimodal Large Language Models, visual token growth, spatial resolution, temporal context, ResAdapt, Input-side adaptation framework, Allocator, Cost-Aware Policy Optimization, contextual bandit, visual budget, video QA, temporal grounding, image reasoning, code available, huggingface daily},
eprint = {2603.28610},
archiveprefix = {arXiv},
}
{}