Seeing Through Words, Speaking Through Pixels: Deep Representational Alignment Between Vision and Language Models

• URL: http://arxiv.org/abs/2509.20751v1
• Date: Thu, 25 Sep 2025 05:16:28 GMT
• Title: Seeing Through Words, Speaking Through Pixels: Deep Representational Alignment Between Vision and Language Models
• Authors: Zoe Wanying He, Sean Trott, Meenakshi Khosla,
• Abstract summary: We find that alignment peaks in mid-to-late layers of both model types.<n>Human preferences for image-caption matches are mirrored in the embedding spaces across all vision-language model pairs.
• Score: 4.5497948012757865
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Recent studies show that deep vision-only and language-only models--trained on disjoint modalities--nonetheless project their inputs into a partially aligned representational space. Yet we still lack a clear picture of where in each network this convergence emerges, what visual or linguistic cues support it, whether it captures human preferences in many-to-many image-text scenarios, and how aggregating exemplars of the same concept affects alignment. Here, we systematically investigate these questions. We find that alignment peaks in mid-to-late layers of both model types, reflecting a shift from modality-specific to conceptually shared representations. This alignment is robust to appearance-only changes but collapses when semantics are altered (e.g., object removal or word-order scrambling), highlighting that the shared code is truly semantic. Moving beyond the one-to-one image-caption paradigm, a forced-choice "Pick-a-Pic" task shows that human preferences for image-caption matches are mirrored in the embedding spaces across all vision-language model pairs. This pattern holds bidirectionally when multiple captions correspond to a single image, demonstrating that models capture fine-grained semantic distinctions akin to human judgments. Surprisingly, averaging embeddings across exemplars amplifies alignment rather than blurring detail. Together, our results demonstrate that unimodal networks converge on a shared semantic code that aligns with human judgments and strengthens with exemplar aggregation.

Related papers

• UniModel: A Visual-Only Framework for Unified Multimodal Understanding and Generation [51.31795451147935]
  We present a unified generative model that supports visual understanding and visual generation within a single pixel-to-pixel diffusion framework.<n>Our goal is to achieve unification along three axes: the model, the tasks, and the representations.<n> Experiments on text-to-image synthesis and image-to-text understanding demonstrate strong cross-modal alignment.
  arXiv  Detail & Related papers  (2025-11-21T03:02:10Z)
• Mind-the-Glitch: Visual Correspondence for Detecting Inconsistencies in Subject-Driven Generation [120.23172120151821]
  We propose a novel approach for disentangling visual and semantic features from the backbones of pre-trained diffusion models.<n>We introduce an automated pipeline that constructs image pairs with annotated semantic and visual correspondences.<n>We propose a new metric, Visual Semantic Matching, that quantifies visual inconsistencies in subject-driven image generation.
  arXiv  Detail & Related papers  (2025-09-26T07:11:55Z)
• Embedding and Enriching Explicit Semantics for Visible-Infrared Person Re-Identification [31.011118085494942]
  Visible-infrared person re-identification (VIReID) retrieves pedestrian images with the same identity across different modalities.<n>Existing methods learn visual content solely from images, lacking the capability to sense high-level semantics.<n>We propose an Embedding and Enriching Explicit Semantics framework to learn semantically rich cross-modality pedestrian representations.
  arXiv  Detail & Related papers  (2024-12-11T14:27:30Z)
• Distractors-Immune Representation Learning with Cross-modal Contrastive Regularization for Change Captioning [71.14084801851381]
  Change captioning aims to succinctly describe the semantic change between a pair of similar images. Most existing methods directly capture the difference between them, which risk obtaining error-prone difference features. We propose a distractors-immune representation learning network that correlates the corresponding channels of two image representations.
  arXiv  Detail & Related papers  (2024-07-16T13:00:33Z)
• Identifying Interpretable Subspaces in Image Representations [54.821222487956355]
  We propose a framework to explain features of image representations using Contrasting Concepts (FALCON) For a target feature, FALCON captions its highly activating cropped images using a large captioning dataset and a pre-trained vision-language model like CLIP. Each word among the captions is scored and ranked leading to a small number of shared, human-understandable concepts.
  arXiv  Detail & Related papers  (2023-07-20T00:02:24Z)
• Learning to Model Multimodal Semantic Alignment for Story Visualization [58.16484259508973]
  Story visualization aims to generate a sequence of images to narrate each sentence in a multi-sentence story. Current works face the problem of semantic misalignment because of their fixed architecture and diversity of input modalities. We explore the semantic alignment between text and image representations by learning to match their semantic levels in the GAN-based generative model.
  arXiv  Detail & Related papers  (2022-11-14T11:41:44Z)
• Single-Stream Multi-Level Alignment for Vision-Language Pretraining [103.09776737512078]
  We propose a single stream model that aligns the modalities at multiple levels. We achieve this using two novel tasks: symmetric cross-modality reconstruction and a pseudo-labeled key word prediction. We demonstrate top performance on a set of Vision-Language downstream tasks such as zero-shot/fine-tuned image/text retrieval, referring expression, and VQA.
  arXiv  Detail & Related papers  (2022-03-27T21:16:10Z)
• Step-Wise Hierarchical Alignment Network for Image-Text Matching [29.07229472373576]
  We propose a step-wise hierarchical alignment network (SHAN) that decomposes image-text matching into multi-step cross-modal reasoning process. Specifically, we first achieve local-to-local alignment at fragment level, following by performing global-to-local and global-to-global alignment at context level sequentially.
  arXiv  Detail & Related papers  (2021-06-11T17:05:56Z)
• Cross-domain Correspondence Learning for Exemplar-based Image Translation [59.35767271091425]
  We present a framework for exemplar-based image translation, which synthesizes a photo-realistic image from the input in a distinct domain. The output has the style (e.g., color, texture) in consistency with the semantically corresponding objects in the exemplar. We show that our method is superior to state-of-the-art methods in terms of image quality significantly.
  arXiv  Detail & Related papers  (2020-04-12T09:10:57Z)

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.