paper-with-me

홈 › Papers

Visual Anchors Are Strong Information Aggregators For Multimodal Large Language Model

2024-05-28 · Haogeng Liu, Quanzeng You, Xiaotian Han, Yongfei Liu, Huaibo Huang, Ran He, Hongxia Yang

In the realm of Multimodal Large Language Models (MLLMs), vision-language connector plays a crucial role to link the pre-trained vision encoders with Large Language Models (LLMs). Despite its importance, the vision-language connector has been relatively less explored. In this study, we aim to propose a strong vision-language connector that enables MLLMs to achieve high accuracy while maintain low computation cost. We first reveal the existence of the visual anchors in Vision Transformer and propose a cost-effective search algorithm to extract them. Building on these findings, we introduce the Anchor Former (AcFormer), a novel vision-language connector designed to leverage the rich prior knowledge obtained from these visual anchors during pretraining, guiding the aggregation of information. Through extensive experimentation, we demonstrate that the proposed method significantly reduces computational costs by nearly two-thirds compared with baseline, while simultaneously outperforming baseline methods. This highlights the effectiveness and efficiency of AcFormer. Codes are available at https://github.com/liuhaogeng/Anchor-Former.

📄 PDF Abstract BibTeX arXiv:2405.17815

Code (1)

liuhaogeng/anchor-former 공식 구현 pytorch

Tasks

Language ModelingLanguage ModellingLarge Language ModelMultimodal Large Language Model

Methods 이 논문이 사용한 방법론

Attention 설명 없음
BPE Byte Pair Encoding, or BPE, is a subword segmentation algorithm that encodes rare and unknown words as sequences of subword units. The intuition is that various word…
Label Smoothing Label Smoothing is a regularization technique that introduces noise for the labels. This accounts for the fact that datasets may have mistakes in them, so maximizing the…
Adam 설명 없음
Position-Wise Feed-Forward Layer 설명 없음
Dropout Dropout is a regularization technique for neural networks that drops a unit (along with connections) at training time with a specified probability $p$ (a common value is…
Dense Connections Dense Connections, or Fully Connected Connections, are a type of layer in a deep neural network that use a linear operation where every input is connected to every output…
Absolute Position Encodings Absolute Position Encodings are a type of position embeddings for [Transformer-based models] where positional encodings are…

Similar Papers 제목 키워드 기반

Reflection Anchors for Propagation-Aware Visual Retention in Long-Chain Multimodal Reasoning

2026-05-10 · Xuan Gong, Hanbo Huang, Hao Zheng, Yiran Zhang 외 arxiv

Long chain-of-thought (CoT) reasoning improves large vision--language models, but visual information often fades during generation, limiting long-horizon multimodal reasoning. Existing methods either re-inject vision at …

Multimodal Reasoning

Credit Where It is Due: Cross-Modality Connectivity Drives Precise Reinforcement Learning for MLLM Reasoning

2026-02-12 · Zhengbo Jiao, Shaobo Wang, Zifan Zhang, Wei Wang 외 arxiv

Reinforcement Learning with Verifiable Rewards (RLVR) has significantly advanced the reasoning capabilities of Multimodal Large Language Models (MLLMs), yet how visual evidence is integrated during reasoning remains poor…

Reinforcement LearningVisual Grounding

Visual Attention Drifts,but Anchors Hold:Mitigating Hallucination in Multimodal Large Language Models via Cross-Layer Visual Anchors

2026-03-26 · Chengxu Yang, Jingling Yuan, Chuang Hu, Jiawei Jiang arxiv

Multimodal Large Language Models often suffer from object hallucination. While existing research utilizes attention enhancement and visual retracing, we find these works lack sufficient interpretability regarding attenti…

SearchEyes: Towards Frontier Multimodal Deep Search Intelligence via Search World Simulation

2026-07-07 · Zhengbo Jiao, Yiming Cheng, Yilei Jiang, Kaituo Feng 외 arxiv

Training multimodal search agents to perform multi-hop reasoning remains challenging due to a fundamental structural disconnect: existing pipelines construct training data, search environments, and reward signals indepen…

Every Picture Tells a Dangerous Story: Memory-Augmented Multi-Agent Jailbreak Attacks on VLMs

2026-04-14 · Jianhao Chen, Haoyang Chen, Hanjie Zhao, Haozhe Liang 외 arxiv

Vision-Language Models (VLMs) expand the attack surface of safety-aligned systems by coupling visual perception with text generation. Existing multimodal jailbreak attacks primarily rely on crafted visual content, advers…

Adversarial Attack