Beyond Cross-Modal Alignment: Measuring and Leveraging Modality Gap in Vision-Language Models (2026.findings-acl)
Copied to clipboard
| Challenge: | a recent study shows that vision-language models have modality gaps that persist even in well-aligned models. |
| Approach: | They propose a modality-dominance score to measure and leverage modality gaps . they propose automatic interpretability metrics to evaluate these features in a scalable manner . |
| Outcome: | The proposed framework allows for training-free probing and editing methods for understanding model perception across genders and generating adversarial examples. |
Similar Papers
Cross-Modal Attribute Insertions for Assessing the Robustness of Vision-and-Language Learning (2023.acl-long)
Copied to clipboard
| Challenge: | Existing approaches to model multimodal data do not leverage cross-modal information . augmenting input text using cross-module attribute insertions results in poor performance . |
| Approach: | They propose a multimodal deep learning approach that adds visual attributes to inputs to enhance model robustness. |
| Outcome: | The proposed approach is modular, controllable, and task-agnostic. |
Vision-Language Models Struggle to Align Entities across Modalities (2025.findings-acl)
Copied to clipboard
| Challenge: | Several real-world applications require the ability to perform cross-modal entity linking . cross-functional entity linking is a skill needed for multimodal code generation and scene understanding . |
| Approach: | They propose a task and benchmark to evaluate cross-modal entity linking performance . they use visual scenes aligned with their textual representations to evaluate performance a question-answering task . |
| Outcome: | The proposed task and benchmark aims to improve cross-modal entity linking performance . it evaluates state-of-the-art vision-language models and humans on the task . |
Probing Cross-modal Semantics Alignment Capability from the Textual Perspective (2022.findings-emnlp)
Copied to clipboard
| Challenge: | In recent years, vision and language pre-training (VLP) models have advanced the state-of-the-art results in a variety of cross-modal downstream tasks. |
| Approach: | They propose a new probing method that is based on image captioning to first empirically study the cross-modal semantics alignment of VLP models. |
| Outcome: | The proposed method analyzes captions generated by five popular VLP models to reveal how well they align with visual words and how well these align with images. |
Vision-and-Language or Vision-for-Language? On Cross-Modal Influence in Multimodal Transformers (2021.emnlp-main)
Copied to clipboard
| Challenge: | Pretrained vision-and-language BERTs aim to learn representations that combine information from both modalities. |
| Approach: | They propose a diagnostic method based on cross-modal input ablation to assess the extent to which pretrained models integrate cross-module information. |
| Outcome: | The proposed method evaluates the model's performance on the other modality based on inputs from one or both modality. |
Seeing Through Words, Speaking Through Pixels: Deep Representational Alignment Between Vision and Language Models (2025.emnlp-main)
Copied to clipboard
| Challenge: | Recent studies show that deep vision-only and language-only models project inputs into a partially aligned representational space. |
| Approach: | They investigate whether a model's representational code is semantically shared . they find that alignment peaks in mid-to-late layers of both model types . |
| Outcome: | a forced-choice "Pick-a-Pic" task shows human preferences for image-caption matches are mirrored in embedding spaces across vision-language model pairs. |
Understanding the Modality Gap: An Empirical Study on the Speech-Text Alignment Mechanism of Large Speech Language Models (2025.emnlp-main)
Copied to clipboard
| Challenge: | LSLMs have impressive conversational generation abilities, but consistently fall short of traditional pipeline systems on semantic understanding benchmarks. |
| Approach: | They propose to analyze the performance gap between speech and text inputs through a systematic experiment . they find that representation similarity is strongly correlated with the modality gap . |
| Outcome: | The proposed models improve the accuracy of speech inputs and their semantic understanding benchmarks. |
MMAC: A Multilingual, Multimodal Alignment Framework for Cultural Grounding Evaluation (2026.acl-long)
Copied to clipboard
Weihua Zheng, Zhengyuan Liu, Tanmoy Chakraborty, Weiwen Xu, Xiaoxue Gao, Bryan Chen Zhengyu Tan, Bowei Zou, Chang Liu, Yujia Hu, Xing Xie, Xiaoyuan Yi, Jing Yao, Chaojun Wang, Long Li, Rui Liu, Huiyao Liu, Koji Inoue, Ryuichi Sumida, Tatsuya Kawahara, Fan Xu, Lingyu Ye, Wei Tian, Dongjun Kim, Jimin Jung, Jaehyung Seo, Nadya Yuki Wangsajaya, Pham Minh Duc, Ojasva Saxena, Palash Nandi, Xiyan Tao, Wiwik Karlina, Tuan Luong, Keertana Arun Vasan, Roy Ka-Wei Lee, Nancy F. Chen
| Challenge: | Existing models lack cultural alignment across modalities and languages . a new framework to assess cultural awareness across linguistics and languages is needed . |
| Approach: | They propose a framework that integrates tri-modally aligned cultural benchmarks and a five-dimensional evaluation protocol to assess cross-country awareness disparities. |
| Outcome: | The proposed framework assesses cultural awareness disparities across modalities and languages . it is the first dataset aligned at the input level across text, image, and speech . |
Understanding and Bridging the Modality Gap for Speech Translation (2023.acl-long)
Copied to clipboard
| Challenge: | Existing methods to improve end-to-end speech translation (ST) use multitask learning, but there is always a modality gap between ST and MT due to the differences between speech and text. |
| Approach: | They propose a method to bridge the modality gap between ST and MT by leveraging (text) machine translation data. |
| Outcome: | The proposed method bridges the modality gap and achieves significant improvements over baseline in all eight directions. |
Multimodal Routing: Improving Local and Global Interpretability of Multimodal Language Analysis (2020.emnlp-main)
Copied to clipboard
| Challenge: | Recent multimodal learning models with strong performances on human-centric tasks are often black-box with very limited interpretability. |
| Approach: | They propose a multimodal routing algorithm which dynamically adjusts weights between input and output modalities for each input sample. |
| Outcome: | The proposed model can interpret modality-prediction relationships globally and locally for each input sample while keeping competitive performance compared to state-of-the-art methods. |
Modality Alignment between Deep Representations for Effective Video-and-Language Learning (2022.lrec-1)
Copied to clipboard
| Challenge: | Existing Video-and-Language models do not take into account the different characteristics of video and text representations. |
| Approach: | They propose a method that exploits Centered Kernel Alignment (CKA) to enhance cross-modality attention by combining multiple modalities. |
| Outcome: | The proposed method outperforms conventional multi-modal methods significantly on video QA tasks with +3.57% accuracy increment compared to the baseline in a popular benchmark dataset. |