paper-with-me

Papers

Connect, Collapse, Corrupt: Learning Cross-Modal Tasks with Uni-Modal Data

2024-01-16 · Yuhui Zhang, Elaine Sui, Serena Yeung-Levy

Building cross-modal applications is challenging due to limited paired multi-modal data. Recent works have shown that leveraging a pre-trained multi-modal contrastive representation space enables cross-modal tasks to be learned from uni-modal data. This is based on the assumption that contrastive optimization makes embeddings from different modalities interchangeable. However, this assumption is under-explored due to the poorly understood geometry of the multi-modal contrastive space, where a modality gap exists. In our study, we provide a theoretical explanation of this space's geometry and introduce a three-step method, $C^3$ (Connect, Collapse, Corrupt), to bridge the modality gap, enhancing the interchangeability of embeddings. Our $C^3$ method significantly improves cross-modal learning from uni-modal data, achieving state-of-the-art results on zero-shot image / audio / video captioning and text-to-image generation.

📄 PDF Abstract BibTeX arXiv:2401.08567

Code (1)

yuhui-zh15/c3 공식 구현 pytorch

Tasks

Image GenerationText to Image GenerationText-to-Image GenerationVideo Captioning

Similar Papers 제목 키워드 기반

Can BEV Perception Gracefully Degrade under Sensor Failures?

2026-05-29 · Haifa Zhang, Yijing Wang, Haoyu Wang, Zheng Li 외 arxiv

Despite the remarkable success of multi-modal bird's-eye view (BEV) perception in autonomous driving, current systems exhibit a critical vulnerability: existing fusion mechanisms are highly brittle to sensor corruptions,…

Autonomous Driving

Multimodal Evaluator Preference Collapse: Cross-Modal Coupling in Self-Evolving Agents

2026-06-15 · Zewen Liu arxiv

When AI agents use language models to evaluate their own outputs in a feedback loop, systematic biases emerge. We show that Evaluator Preference Collapse (EPC) is dramatically amplified in multimodal settings. Using GPT-…

Robo-MUTUAL: Robotic Multimodal Task Specification via Unimodal Learning

2024-10-02 · Jianxiong Li, Zhihao Wang, Jinliang Zheng, Xiaoai Zhou 외

Multimodal task specification is essential for enhanced robotic performance, where \textit{Cross-modality Alignment} enables the robot to holistically understand complex task instructions. Directly annotating multimodal …

SplAttN: Bridging 2D and 3D with Gaussian Soft Splatting and Attention for Point Cloud Completion

2026-05-02 · Zhaoyang Li, Zhichao You, Tianrui Li arxiv

Although multi-modal learning has advanced point cloud completion, the theoretical mechanisms remain unclear. Recent works attribute success to the connection between modalities, yet we identify that standard hard projec…

Point Cloud CompletionDensity Estimation

Latent Denoising Improves Visual Alignment in Large Multimodal Models

2026-04-23 · Dhruv Parikh, Jacob Fein-Ashley, Rajgopal Kannan, Viktor Prasanna arxiv

Large Multimodal Models (LMMs) such as LLaVA are typically trained with an autoregressive language modeling objective, providing only indirect supervision to visual tokens. This often yields weak internal visual represen…