MLLMs are Deeply Affected by Modality Bias
read the original abstract
Recent advances in Multimodal Large Language Models (MLLMs) have shown promising results in integrating diverse modalities such as texts and images. MLLMs are heavily influenced by modality bias, often relying on language while under-utilizing other modalities like visual inputs. This position paper argues that MLLMs are deeply affected by modality bias. Firstly, we diagnose the current state of modality bias, highlighting its manifestations across various tasks. Secondly, we propose a systematic research road-map related to modality bias in MLLMs. Thirdly, we identify key factors of modality bias in MLLMs and offer actionable suggestions for future research to mitigate it. To substantiate these findings, we conduct experiments that demonstrate the influence of each factor: 1. Data Characteristics: Language data is compact and abstract, while visual data is redundant and complex, creating an inherent imbalance in learning dynamics. 2. Imbalanced Backbone Capabilities: The dominance of pretrained language models in MLLMs leads to overreliance on language and neglect of visual information. 3. Training Objectives: Current objectives often fail to promote balanced cross-modal alignment, resulting in shortcut learning biased toward language. These findings highlight the need for balanced training strategies and model architectures to better integrate multiple modalities in MLLMs. We call for interdisciplinary efforts to tackle these challenges and drive innovation in MLLM research. Our work provides a fresh perspective on modality bias in MLLMs and offers insights for developing more robust and generalizable multimodal systems-advancing progress toward Artificial General Intelligence.
This paper has not been read by Pith yet.
Forward citations
Cited by 12 Pith papers
-
Sci-Rho: A Multilingual Visually-Grounded Symbolic Benchmark for STEM Problems
Sci-Rho is a dynamic multilingual visually-grounded symbolic benchmark for STEM problems that reveals robustness gaps in current VLMs between average and worst-case performance.
-
Measuring Cross-Modal Synergy: A Benchmark for VLM Explainability
Introduces Synergistic Faithfulness metric based on Shapley Interaction Index to evaluate cross-modal synergy in VLM explainers, revealing over-reliance on visual salience in existing methods.
-
Vesta: A Generalist Embodied Reasoning Model
Vesta is a unified embodied generalist model that outperforms specialist baselines by over 20% on average and improves real-world robotic task success by over 35%.
-
Pareto LoRA: Mitigating Modality Imbalance in Unified Multimodal Models via Pareto-Optimal Gradient Integration
Pareto LoRA applies Pareto-optimal gradient integration to balance text and image objectives in LoRA-based fine-tuning of unified multimodal models, reporting up to 44.9% gains in image quality on the CoMM benchmark w...
-
Semantic Generative Tuning for Unified Multimodal Models
Semantic Generative Tuning uses image segmentation as a generative proxy to align misaligned representation spaces in unified multimodal models and improve both perception and generative layout fidelity.
-
MiMIC: Mitigating Visual Modality Collapse in Universal Multimodal Retrieval While Avoiding Semantic Misalignment
MiMIC mitigates visual modality collapse and semantic misalignment in universal multimodal retrieval via fusion-in-decoder architecture and robust single-modality training.
-
Beyond Text-Dominance: Understanding Modality Preference of Omni-modal Large Language Models
Omni-modal LLMs exhibit visual preference that emerges in mid-to-late layers, enabling hallucination detection without task-specific training.
-
A Systematic Study of Cross-Modal Typographic Attacks on Audio-Visual Reasoning
Coordinated multi-modal typographic attacks on MLLMs achieve 83.43% success rate versus 34.93% for single-modality attacks.
-
Token-Efficient Multimodal Reasoning via Image Prompt Packaging
IPPg embeds text into images to reduce multimodal model inference costs by 35.8-91% with competitive accuracy on many VQA and code benchmarks.
-
CAAD: Contrastive Audio-Aware Distillation for Efficient Speech Language Models
CAAD internalizes contrastive audio-aware decoding into student SLM weights via synchronized teacher-forcing, delivering an 8% relative gain over standard knowledge distillation on Dynamic-SUPERB while reducing lingui...
-
Semantic Generative Tuning for Unified Multimodal Models
Semantic Generative Tuning applies segmentation-based generative proxies during post-training to align and improve both understanding and generation in unified multimodal models.
-
Do Composed Image Retrieval Benchmarks Require Multimodal Composition?
CIR benchmarks contain many unimodal shortcuts and noisy queries, leading to overestimation of models' multimodal composition capabilities.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.