Research map / Vision-language and multimodal / Human and embodied modeling
Aspect-Aware Multimodal Summarization for Chinese E-Commerce Products
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 |
|---|---|---|---|
| add global visual feature to decoder initialization (MMPG+DecInit) (ablation) | pointer-generator · CEPSUM Home Appliances | ROUGE-1: 31.31 -> 32.88 | Helped |
| add global visual feature to decoder initialization (ablation) | pointer-generator · CEPSUM Clothing | ROUGE-1: 29.11 -> 30.73 | Helped |
| initialize encoder instead of decoder with visual feature (ablation) | MMPG · CEPSUM Home Appliances | ROUGE-1: 32.88 -> 32.67 | Hurt |
| add local-feature multimodal attention to DecInit (ablation) | MMPG+DecInit · CEPSUM Home Appliances | ROUGE-1: 32.88 -> 32.87 | No clear effect |
| add aspect strategies (RAML, coverage, coherence) to MMPG (ablation) | MMPG · CEPSUM Home Appliances | ROUGE-1: 32.88 -> 33.97 | Helped |
| add consistency constraint on attribute words (ablation) | MMPG+Aspect · CEPSUM Home Appliances | ROUGE-1: 33.97 -> 34.36 | Helped |
| add aspect RAML training (ablation) | MMPG · CEPSUM Home Appliances manual eval | aspect importance (%): 78.00 -> 91.00 | Helped |
| add aspect RAML training (ablation) | MMPG · CEPSUM Home Appliances manual eval | readability: 3.05 -> 3.06 | No clear effect |
| add aspect coverage (ablation) | MMPG+AspRAML · CEPSUM Clothing manual eval | readability: 3.31 -> 3.22 | Hurt |
| add aspect coherence (constrained decoding) (ablation) | MMPG+AspRAML+AspCov · CEPSUM Clothing manual eval | readability: 3.22 -> 3.64 | Helped |
| add consistency constraint blocking generated attribute words (ablation) | MMPG+Aspect · CEPSUM Home Appliances manual eval | consistent summaries (%): 71 -> 92 | Helped |
| add aspect coverage mechanism (ablation) | MMPG · CEPSUM | repeated words and aspects | Helped |
Closest papers by meaning
- Tell Me What Is Good about This Property: Leveraging Reviews for Segment-Personalized Image Collection Summarization (AAAI 2024 · similarity 0.84)
- BizGen: Advancing Article-level Visual Text Rendering for Infographics Generation (CVPR 2025 · similarity 0.82)
- DIUSum: Dynamic Image Utilization for Multimodal Summarization (AAAI 2024 · similarity 0.81)
- TIGER-FG: Text-Guided Implicit Fine-Grained Grounding for E-commerce Retrieval (NeurIPS 2026 · similarity 0.81)
- QuerySum: A Multi-Document Query-Focused Summarization Dataset Augmented with Similar Query Clusters (AAAI 2024 · similarity 0.81)
- MultiSum: A Multi-Facet Approach for Extractive Social Summarization Utilizing Semantic and Sociological Relationships (AAAI 2024 · similarity 0.81)
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.