ToVE: Efficient Vision-Language Learning via Knowledge Transfer from Vision Experts

• URL: http://arxiv.org/abs/2504.00691v1
• Date: Tue, 01 Apr 2025 12:02:40 GMT
• Title: ToVE: Efficient Vision-Language Learning via Knowledge Transfer from Vision Experts
• Authors: Yuanchen Wu, Junlong Du, Ke Yan, Shouhong Ding, Xiaoqiang Li,
• Abstract summary: Vision-language (VL) learning requires extensive visual perception capabilities.<n>Recent works typically rely on training huge models on massive datasets to develop these capabilities.<n>This paper proposes a new framework that transfers the knowledge from a hub of Vision Experts.
• Score: 29.446235941754345
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Vision-language (VL) learning requires extensive visual perception capabilities, such as fine-grained object recognition and spatial perception. Recent works typically rely on training huge models on massive datasets to develop these capabilities. As a more efficient alternative, this paper proposes a new framework that Transfers the knowledge from a hub of Vision Experts (ToVE) for efficient VL learning, leveraging pre-trained vision expert models to promote visual perception capability. Specifically, building on a frozen CLIP encoder that provides vision tokens for image-conditioned language generation, ToVE introduces a hub of multiple vision experts and a token-aware gating network that dynamically routes expert knowledge to vision tokens. In the transfer phase, we propose a "residual knowledge transfer" strategy, which not only preserves the generalizability of the vision tokens but also allows detachment of low-contributing experts to improve inference efficiency. Further, we explore to merge these expert knowledge to a single CLIP encoder, creating a knowledge-merged CLIP that produces more informative vision tokens without expert inference during deployment. Experiment results across various VL tasks demonstrate that the proposed ToVE achieves competitive performance with two orders of magnitude fewer training data.

Related papers

• Expanding the Boundaries of Vision Prior Knowledge in Multi-modal Large Language Models [53.13731845500678]
  We introduce a novel metric, $Rank_e$, to quantify the effect of vision encoder's prior knowledge on MLLM performance.<n>We propose VisPRE, a two-stage training framework that explicitly incorporates prior knowledge at the vision encoder level.<n> Experimental results demonstrate that augmenting vision encoder's prior knowledge substantially boosts the visual understanding capabilities of MLLMs.
  arXiv  Detail & Related papers  (2025-03-23T11:33:09Z)
• Beyond Sight: Towards Cognitive Alignment in LVLM via Enriched Visual Knowledge [24.538839144639653]
  Large Vision-Language Models (LVLMs) integrate separately pre-trained vision and language components. These models frequently encounter a core issue of "cognitive misalignment" between the vision encoder (VE) and the large language model (LLM)
  arXiv  Detail & Related papers  (2024-11-25T18:33:14Z)
• VLM-KD: Knowledge Distillation from VLM for Long-Tail Visual Recognition [25.927771583678272]
  We introduce an effective method to distill knowledge from an off-the-shelf vision-language model (VLM) We develop a framework that generates novel text supervision and distills free-form text into a vision encoder. To our knowledge, this work is the first to utilize text supervision generated by an off-the-shelf VLM and apply it to vanilla randomly vision encoders.
  arXiv  Detail & Related papers  (2024-08-29T22:13:29Z)
• Unveiling Encoder-Free Vision-Language Models [62.52803514667452]
  Existing vision-language models (VLMs) mostly rely on vision encoders to extract visual features followed by large language models (LLMs) for visual-language tasks. We bridge the gap between encoder-based and encoder-free models, and present a simple yet effective training recipe towards pure VLMs. We launch EVE, an encoder-free vision-language model that can be trained and forwarded efficiently.
  arXiv  Detail & Related papers  (2024-06-17T17:59:44Z)
• MoVA: Adapting Mixture of Vision Experts to Multimodal Context [38.8308841469793]
  We propose the MoVA, a powerful and novel MLLM, adaptively routing and fusing task-specific vision experts with a coarse-to-fine mechanism. In the coarse-grained stage, we design a context-aware expert routing strategy to dynamically select the most suitable vision experts. In the fine-grained stage, we elaborately conduct the mixture-of-vision-expert adapter (MoV-Adapter) to extract and fuse task-specific knowledge.
  arXiv  Detail & Related papers  (2024-04-19T17:59:48Z)
• Incorporating Visual Experts to Resolve the Information Loss in Multimodal Large Language Models [121.83413400686139]
  This paper proposes to improve the visual perception ability of MLLMs through a mixture-of-experts knowledge enhancement mechanism. We introduce a novel method that incorporates multi-task encoders and visual tools into the existing MLLMs training and inference pipeline.
  arXiv  Detail & Related papers  (2024-01-06T02:02:34Z)
• A Comprehensive Evaluation of GPT-4V on Knowledge-Intensive Visual Question Answering [53.70661720114377]
  multimodal large models (MLMs) have significantly advanced the field of visual understanding, offering remarkable capabilities in realm of visual question answering (VQA) Yet, the true challenge lies in the domain of knowledge-intensive VQA tasks, which necessitate deep comprehension of the visual information in conjunction with a vast repository of learned knowledge. To uncover such capabilities, we provide an in-depth evaluation from three perspectives: 1) Commonsense Knowledge, which assesses how well models can understand visual cues and connect to general knowledge; 2) Fine-grained World Knowledge, which tests the model's skill in reasoning out specific knowledge from images, showcasing
  arXiv  Detail & Related papers  (2023-11-13T18:22:32Z)
• K-LITE: Learning Transferable Visual Models with External Knowledge [242.3887854728843]
  K-LITE (Knowledge-augmented Language-Image Training and Evaluation) is a strategy to leverage external knowledge to build transferable visual systems. In training, it enriches entities in natural language with WordNet and Wiktionary knowledge. In evaluation, the natural language is also augmented with external knowledge and then used to reference learned visual concepts.
  arXiv  Detail & Related papers  (2022-04-20T04:47:01Z)
• Reasoning over Vision and Language: Exploring the Benefits of Supplemental Knowledge [59.87823082513752]
  This paper investigates the injection of knowledge from general-purpose knowledge bases (KBs) into vision-and-language transformers. We empirically study the relevance of various KBs to multiple tasks and benchmarks. The technique is model-agnostic and can expand the applicability of any vision-and-language transformer with minimal computational overhead.
  arXiv  Detail & Related papers  (2021-01-15T08:37:55Z)

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.