Research map / Vision-language and multimodal / Human and embodied modeling
Deliberate Attention Networks for Image Captioning
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 deliberate second-pass attention layer (without RL) (ablation) | DA · COCO ablation | CIDEr: 104.9 -> 111.9 | Helped |
| add deliberate second-pass attention layer (with RL) (ablation) | DA · COCO ablation | CIDEr: 117.3 -> 125.6 | Helped |
| add reinforcement learning with CIDEr and contrastive reward (without deliberation) (ablation) | DA · COCO ablation | CIDEr: 104.9 -> 117.3 | Helped |
| add reinforcement learning with CIDEr and contrastive reward (with deliberation) (ablation) | DA · COCO ablation | CIDEr: 111.9 -> 125.6 | Helped |
| add deliberation (ablation) | DA · COCO ablation | BLEU-4: 33.7 -> 35.7 | Helped |
| add deliberation and RL (full DA) versus Up-Down | COCO Karpathy test · COCO Karpathy test | CIDEr: 120.1 -> 125.6 | Helped |
| DA single model versus Up-Down 4-model ensemble | MSCOCO online server · MSCOCO online c5 | BLEU-4: 36.9 -> 36.8 BLEU-4; 117.9 -> 120.5 CIDEr | Mixed |
Closest papers by meaning
- Text-Guided Attention Model for Image Captioning (AAAI 2017 · similarity 0.89)
- Knowing When to Look: Adaptive Attention via a Visual Sentinel for Image Captioning (CVPR 2017 · similarity 0.89)
- Look Back and Predict Forward in Image Captioning (CVPR 2019 · similarity 0.89)
- Visual Attention Never Fades: Selective Progressive Attention ReCalibration for Detailed Image Captioning in Multimodal Large Language Models (ICML 2025 · similarity 0.88)
- Painting with Words: Elevating Detailed Image Captioning with Benchmark and Alignment Learning (ICLR 2025 · similarity 0.87)
- CapRL: Stimulating Dense Image Caption Capabilities via Reinforcement Learning (ICLR 2026 · similarity 0.87)
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.