Research map / Vision-language and multimodal / Human and embodied modeling
Uncertainty-Aware Image Captioning
Zhengcong Fei, Mingyuan Fan, Li Zhu, Junshi Huang, Xiaoming Wei, Xiaolin Wei
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 |
|---|---|---|---|
| generate words in increasing uncertainty order rather than sequentially (ablation) | UAIC · MS COCO test | CIDEr: 129.2 -> 131.7 | Improved |
| generate words in increasing uncertainty order rather than randomly (ablation) | UAIC · MS COCO test | CIDEr: 129.6 -> 131.7 | Improved |
| generate easy words before complex words instead of using anti-uncertainty order (ablation) | UAIC · MS COCO test | CIDEr: 127.5 -> 131.7 | Improved |
| use uncertainty-adaptive beam search rather than fixed beam size 1 (ablation) | UAIC · MS COCO test | CIDEr: 128.5 -> 131.7 | Improved |
| use uncertainty-adaptive beam search rather than fixed beam size 3 (ablation) | UAIC · MS COCO test | Latency: 52ms -> 55ms | Hurt |
| use uncertainty-adaptive beam search rather than fixed beam size 5 (ablation) | UAIC · MS COCO test | Latency: 77ms -> 55ms | Improved |
| use uncertainty-adaptive beam search rather than fixed beam size 5 (ablation) | UAIC · MS COCO test | CIDEr: 132.1 -> 131.7 | Hurt |
Closest papers by meaning
- Noise-Aware Image Captioning with Progressively Exploring Mismatched Words (AAAI 2024 · similarity 0.90)
- Hierarchical Multi-Source Uncertainty Aggregation for Interactive Video Captioning (AAAI 2025 · similarity 0.89)
- Towards Understanding and Quantifying Uncertainty for Text-to-Image Generation (CVPR 2025 · similarity 0.89)
- Learning Descriptive Image Captioning via Semipermeable Maximum Likelihood Estimation (NeurIPS 2023 · similarity 0.88)
- Noise-Aware Learning from Web-Crawled Image-Text Data for Image Captioning (ICCV 2023 · similarity 0.88)
- Top-Down Semantic Refinement for Image Captioning (AAAI 2026 · 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.
