Research map / Vision-language and multimodal / Cross-modal learning
Structures Meet Semantics: Multimodal Fusion via Graph Contrastive Learning
Jiangfeng Sun, SiHao He, Zhonghong Ou, Meina Song
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| add semantic anchor (ablation) | SSU · CMU-MOSI | ACC2: 86.89% -> 89.32% | Improved |
| add semantic anchor (ablation) | SSU · CMU-MOSEI | ACC2: 87.37% -> 87.93% | Improved |
| add semantic anchor (ablation) | SSU · CMU-MOSI | MAE: 0.5990 -> 0.5666 | Improved |
| add multi-view contrastive losses to regression loss (ablation) | SSU · CMU-MOSI | ACC2: 86.45% -> 89.32% | Improved |
| add multi-view contrastive losses to regression loss (ablation) | SSU · CMU-MOSEI | ACC2: 86.15% -> 87.93% | Improved |
| add Lalign on top of Lreg + Lsup + Lself (ablation) | SSU · CMU-MOSI | ACC2: 88.97% -> 89.32% | Improved |
| add Lalign on top of Lreg + Lsup + Lself (ablation) | SSU · CMU-MOSEI | ACC2: 87.83% -> 87.93% | Improved |
| graph constructor: SSU vs LLM-based graph constructors | SSU · CMU-MOSI | ACC2: 87.65% -> 89.32% | Improved |
| graph constructor: SSU vs LLM-based graph constructors | SSU · graph construction time | seconds per batch: 8.45s -> 0.015s | Improved |
Closest papers by meaning
- SemMSA: Latent Semantic-Aided Robust Multimodal Sentiment Analysis with Incomplete Data (NeurIPS 2026 · similarity 0.90)
- Towards Multimodal Sentiment Analysis via Hierarchical Correlation Modeling with Semantic Distribution Constraints (AAAI 2025 · similarity 0.90)
- A Unified Self-Distillation Framework for Multimodal Sentiment Analysis with Uncertain Missing Modalities (AAAI 2024 · similarity 0.90)
- Enriching Multimodal Sentiment Analysis Through Textual Emotional Descriptions of Visual-Audio Content (AAAI 2025 · similarity 0.89)
- Toward Robust Incomplete Multimodal Sentiment Analysis via Hierarchical Representation Learning (NeurIPS 2024 · similarity 0.89)
- DLF: Disentangled-Language-Focused Multimodal Sentiment Analysis (AAAI 2025 · similarity 0.89)
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.
