Research map / Label-efficient and robust learning / Domain adaptation and generalization
OSAN: A One-Stage Alignment Network To Unify Multimodal Alignment and Unsupervised Domain Adaptation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use one-stage TAL and DDG alignment rather than the two-stage MDMN approach | Multimodal UDA · Text-news → Text-show | F1: 88.34 → 90.53 | Helped |
| Use DDG instead of Mixup | OSAN with TAL · Text-news → Text-show | F1: 89.35 → 90.53 | Helped |
| Add DDG to OSAN (ablation) | OSAN · CMU-MOSI | MAE: +0.039 | Helped |
| Add TAL to OSAN (ablation) | OSAN · CMU-MOSI | MAE: +0.051 | Helped |
| Add both TAL and DDG (ablation) | OSAN · CMU-MOSI | MAE: +0.074 | Helped |
| Add adversarial and discrepancy losses to classification loss (ablation) | OSAN · CMU-MOSI | MAE: +0.018 | Helped |
| Add discrepancy loss alongside classification and adversarial losses (ablation) | OSAN · CMU-MOSI | MAE: +0.009 | Helped |
| Add adversarial loss alongside classification and discrepancy losses (ablation) | OSAN · CMU-MOSI | MAE: +0.016 | Helped |
| Set DDG fusion coefficient to 0.85 rather than 0.9 (ablation) | OSAN · CMU-MOSI | MAE: +0.004 | Helped |
| Use DDG multi-head count 64 rather than 32 (ablation) | OSAN · CMU-MOSI | MAE: 0.001 | Helped |
| Add DDG to OSAN (ablation) | OSAN · Text-show | F1: -1.09 | Helped |
| Add TAL to OSAN (ablation) | OSAN · Text-show | F1: -1.55 | Helped |
Closest papers by meaning
- MODfinity: Unsupervised Domain Adaptation with Multimodal Information Flow Intertwining (CVPR 2025 · similarity 0.90)
- UNIALIGN: Scaling Multimodal Alignment within One Unified Model (CVPR 2025 · similarity 0.89)
- With Limited Data for Multimodal Alignment, Let the STRUCTURE Guide You (NeurIPS 2025 · similarity 0.89)
- Strong Helps Weak: Directional Cross-Modal Alignment Transfer in Multi-modal LLMs (NeurIPS 2026 · similarity 0.89)
- Split to Merge: Unifying Separated Modalities for Unsupervised Domain Adaptation (CVPR 2024 · similarity 0.88)
- AlignMamba: Enhancing Multimodal Mamba with Local and Global Cross-modal Alignment (CVPR 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.