Research map / Vision-language and multimodal / Cross-modal learning
Prototype-as-Prompt: Multimodal Sentiment Prototypes Endowing Large Language Models the Capability to Perform Multimodal Sentiment Analysis
Xianbing Zhao, Lan Luo, Hengyang Lu, Buzhou Tang
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 |
|---|---|---|---|
| replace MSE with Prototype-as-Prompt soft prompts | Llama-2-7B frozen LLM · CMU-MOSEI | Acc-2: 86.74 -> 87.17 | Improved |
| replace MSE with Prototype-as-Prompt soft prompts | ChatGLM-6B frozen LLM · SIMS-V2 | Acc-2: 83.77 -> 84.75 | Improved |
| replace MSE with Prototype-as-Prompt soft prompts | Llama-2-7B frozen LLM · SIMS-V2 | Acc-2: 75.53 -> 76.23 | Improved |
| replace MSE with Prototype-as-Prompt soft prompts | ChatGLM-6B frozen LLM · CHERMA | Acc: 72.90 -> 75.33 | Improved |
| replace MSE with PaP | Qwen-1.5B frozen LLM · CMU-MOSEI | Acc-2: 84.12 -> 86.00 | Improved |
| add cross-modal prototype alignment (CMD loss) (ablation) | PaP (ChatGLM-6B) · CHERMA | Acc-2: 74.01 -> 75.33 | Improved |
| add sentiment-aware prototype learning loss (ablation) | PaP (ChatGLM-6B) · CHERMA | Acc-2: 73.83 -> 75.33 | Improved |
| add distance-weighted prototype diversity regularization (ablation) | PaP (ChatGLM-6B) · CHERMA | Acc-2: 74.67 -> 75.33 | Improved |
| add sentiment-aware prototype learning loss (ablation) | PaP (Llama-2-7B) · CHERMA | Acc-2: 71.69 -> 72.70 | Improved |
| add cross-modal prototype alignment (ablation) | PaP (Qwen-1.5B) · CHERMA | Acc-2: 73.00 -> 73.93 | Improved |
| add diversity regularization (ablation) | PaP (Qwen-1.5B) · CHERMA | Acc-2: 73.47 -> 73.93 | Improved |
| choose LLM backbone Llama-2-7B instead of ChatGLM-6B | PaP · SIMS-V2 | Acc-2: 84.75 -> 76.23 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- AlignMamba — AlignMamba: Enhancing Multimodal Mamba with Local and Global Cross-modal Alignment (CVPR 2025) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Enriching Multimodal Sentiment Analysis Through Textual Emotional Descriptions of Visual-Audio Content (AAAI 2025 · similarity 0.87)
- DLF: Disentangled-Language-Focused Multimodal Sentiment Analysis (AAAI 2025 · similarity 0.87)
- SemMSA: Latent Semantic-Aided Robust Multimodal Sentiment Analysis with Incomplete Data (NeurIPS 2026 · similarity 0.87)
- Towards Multimodal Sentiment Analysis via Hierarchical Correlation Modeling with Semantic Distribution Constraints (AAAI 2025 · similarity 0.87)
- Toward Robust Incomplete Multimodal Sentiment Analysis via Hierarchical Representation Learning (NeurIPS 2024 · similarity 0.87)
- An Empirical Study on Configuring In-Context Learning Demonstrations for Unleashing MLLMs' Sentimental Perception Capability (ICML 2025 · similarity 0.86)
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.
