Rethinking Homogeneity of Vision and Text Tokens in Large Vision-and-Language Models

• URL: http://arxiv.org/abs/2502.01906v1
• Date: Tue, 04 Feb 2025 00:46:11 GMT
• Title: Rethinking Homogeneity of Vision and Text Tokens in Large Vision-and-Language Models
• Authors: Chia-Wen Kuo, Sijie Zhu, Fan Chen, Xiaohui Shen, Longyin Wen,
• Abstract summary: We propose De Attention (D-Attn), a novel method that processes visual and textual embeddings differently. D-Attn diagonalizes visual-to-visual self-attention, reducing computation from $mathcalO(|V|2)$ to $mathcalO(|V|)$ for $|V|$ visual embeddings without compromising performance.
• Score: 29.611769371733672
• License:
• Abstract: Large vision-and-language models (LVLMs) typically treat visual and textual embeddings as homogeneous inputs to a large language model (LLM). However, these inputs are inherently different: visual inputs are multi-dimensional and contextually rich, often pre-encoded by models like CLIP, while textual inputs lack this structure. In this paper, we propose Decomposed Attention (D-Attn), a novel method that processes visual and textual embeddings differently by decomposing the 1-D causal self-attention in LVLMs. After the attention decomposition, D-Attn diagonalizes visual-to-visual self-attention, reducing computation from $\mathcal{O}(|V|^2)$ to $\mathcal{O}(|V|)$ for $|V|$ visual embeddings without compromising performance. Moreover, D-Attn debiases positional encodings in textual-to-visual cross-attention, further enhancing visual understanding. Finally, we introduce an $\alpha$-weighting strategy to merge visual and textual information, maximally preserving the pre-trained LLM's capabilities with minimal modifications. Extensive experiments and rigorous analyses validate the effectiveness of D-Attn, demonstrating significant improvements on multiple image benchmarks while significantly reducing computational costs. Code, data, and models will be publicly available.

Related papers

• Symmetrical Visual Contrastive Optimization: Aligning Vision-Language Models with Minimal Contrastive Images [7.823336661261962]
  Large Vision-Language Models (VLMs) tend to neglect image content and over-rely on language-model priors. We propose S-VCO (Symmetrical Visual Contrastive Optimization), a novel finetuning objective that steers the model toward capturing important visual details.
  arXiv  Detail & Related papers  (2025-02-19T18:05:42Z)
• AdaFV: Rethinking of Visual-Language alignment for VLM acceleration [7.9213473377478865]
  Some approaches to reduce the visual tokens according to the self-attention of VLMs, which are biased, result in inaccurate responses. We propose a self-adaptive cross-modality attention mixture mechanism that dynamically leverages the effectiveness of visual saliency and text-to-image similarity. The proposed approach achieves state-of-the-art training-free VLM acceleration performance, especially when the reduction rate is sufficiently large.
  arXiv  Detail & Related papers  (2025-01-16T13:34:33Z)
• Looking Beyond Text: Reducing Language bias in Large Vision-Language Models via Multimodal Dual-Attention and Soft-Image Guidance [67.26434607115392]
  Large vision-language models (LVLMs) have achieved impressive results in various vision-language tasks. LVLMs suffer from hallucinations caused by language bias, leading to diminished focus on images and ineffective visual comprehension. We propose LACING to address the language bias of LVLMs with muLtimodal duAl-attention meChanIsm (MDA) aNd soft-image Guidance (IFG)
  arXiv  Detail & Related papers  (2024-11-21T16:33:30Z)
• Fine-Grained Verifiers: Preference Modeling as Next-token Prediction in Vision-Language Alignment [57.0121616203175]
  We propose FiSAO, a novel self-alignment method that utilizes the model's own visual encoder as a fine-grained verifier to improve vision-language alignment. By leveraging token-level feedback from the vision encoder, FiSAO significantly improves vision-language alignment, even surpassing traditional preference tuning methods that require additional data.
  arXiv  Detail & Related papers  (2024-10-18T03:34:32Z)
• VideoLLM-MoD: Efficient Video-Language Streaming with Mixture-of-Depths Vision Computation [66.00245701441547]
  We introduce a novel approach to reduce vision compute by leveraging redundant vision tokens "skipping layers" rather than decreasing the number of vision tokens. Our method, VideoLLM-MoD, is inspired by mixture-of-depths LLMs and addresses the challenge of numerous vision tokens in long-term or streaming video.
  arXiv  Detail & Related papers  (2024-08-29T17:21:58Z)
• Visually Descriptive Language Model for Vector Graphics Reasoning [76.42082386029206]
  We propose the Visually Descriptive Language Model (VDLM) to bridge the gap between low-level visual perception and high-level language reasoning. We show that VDLM significantly improves state-of-the-art LMMs like GPT-4o on various multimodal perception and reasoning tasks.
  arXiv  Detail & Related papers  (2024-04-09T17:30:18Z)
• Unified Language-Vision Pretraining in LLM with Dynamic Discrete Visual Tokenization [52.935150075484074]
  We introduce a well-designed visual tokenizer to translate the non-linguistic image into a sequence of discrete tokens like a foreign language. The resulting visual tokens encompass high-level semantics worthy of a word and also support dynamic sequence length varying from the image. This unification empowers LaVIT to serve as an impressive generalist interface to understand and generate multi-modal content simultaneously.
  arXiv  Detail & Related papers  (2023-09-09T03:01:38Z)
• DiMBERT: Learning Vision-Language Grounded Representations with Disentangled Multimodal-Attention [101.99313208598569]
  Vision-and-language (V-L) tasks require the system to understand both vision content and natural language. We propose DiMBERT (short for Disentangled Multimodal-Attention BERT), which applies separated attention spaces for vision and language. We show that DiMBERT sets new state-of-the-art performance on three tasks.
  arXiv  Detail & Related papers  (2022-10-28T23:00:40Z)
• VLMAE: Vision-Language Masked Autoencoder [21.97700040013084]
  We propose a vision-language masked autoencoder framework (VLMAE) for vision-language pre-training. VLMAE employs visual generative learning, facilitating the model to acquire fine-grained and unbiased features.
  arXiv  Detail & Related papers  (2022-08-19T14:39:18Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.