The paper introduces MEQ, a mutual feedback architecture that iteratively refines two multimodal inputs into coupled embeddings, each embedding incorporating information from the other. By continuously exchanging information between the modalities, the model converges to a fixed point that improves representation quality. Experiments on classification and visual grounding tasks show that MEQ achieves competitive or superior performance compared to concatenation-based baselines, and qualitatively enhances visual grounding when paired with complementary modalities.
This work proposes a mutual feedback architecture, MEQ, that refines the two inputs, of possibly different modalities, into a pair of coupled embeddings such that each embedding reflects the informati...
arXiv:2606. 11614v1 Announce Type: cross Abstract: Multimodal learning hinges on capturing redundant, unique, and synergistic information across modalities, which collectively constitute multimodal interactions.
arXiv:2505. 19614v2 Announce Type: replace Abstract: Multimodal learning has seen remarkable progress, particularly with large-scale pre-training across various modalities.
arXiv:2607. 16789v1 Announce Type: new Abstract: Real-world perception and decision making are inherently multimodal, integrating complementary signals across modalities.
arXiv:2606. 05109v1 Announce Type: new Abstract: To leverage the full potential of multimodal data, we need representations that go beyond the state-of-the-art alignment and fusion approaches and exploit all cross-modal interactions without sacrificing modality-specific information.
arXiv:2609.37225v1 Announce Type: cross Abstract: Multimodal large language models (MLLMs) have shown strong potential for universal multimodal representation learning. However, existing methods eith...
arXiv:2609.40362v1 Announce Type: new Abstract: We present Multimodal Flow, a fully continuous generative model of language and vision. Most unified multimodal models either model both language and q...
arXiv:2610.00576v1 Announce Type: new Abstract: In this paper, we propose Gestalt, a new paradigm of large multimodal model built around multimodal interplay. Despite rapid advances, large multimodal...
arXiv:2506. 08774v2 Announce Type: replace-cross Abstract: Different machine learning models can represent the same underlying concept in different ways.
The paper investigates image tokenizers as the visual language of unified multimodal models by creating a controlled autoregressive testbed that tracks task‑specific validation losses during multimodal continual pretraining across text, image, text‑to‑image, and image‑to‑text predictions. It shows that losses must be analyzed by task, that the loss–performance relationship varies with the token space, and that better reconstruction does not always lead to stronger downstream performance. The study also demonstrates how tokenizer design choices—such as discriminator use, semantic supervision, and vocabulary size—affect joint modeling and downstream results.
To leverage the full potential of multimodal data, we need representations that go beyond the state-of-the-art alignment and fusion approaches and exploit all cross-modal interactions without sacrificing modality-specific information. Learning disentangled representations is a principled way to identify these underlying shared and unique factors that are hidden in observational data.
arXiv:2508. 12466v2 Announce Type: replace-cross Abstract: Traditional multimodal learning approaches rely on alignment pre-training to bridge vision and language modalities, typically by projecting visual features into discrete text token spaces using large-scale image--text data.