Research map / Vision-language and multimodal / Human and embodied modeling
Action-Aware Embedding Enhancement for Image-Text Retrieval
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 |
|---|---|---|---|
| Use action-aware memory-enhanced embeddings with ensemble and pre-trained BERT text embeddings instead of DIME | Image-text retrieval · Flickr30K | Text retrieval R@1: 81.0 -> 81.9 | Helped |
| Use action-aware memory-enhanced embeddings with ensemble and pre-trained BERT text embeddings instead of DIME | Image-text retrieval · Microsoft COCO (1K) | Text retrieval R@1: 78.8 -> 79.4 | Helped |
| Use action-aware memory-enhanced embeddings with ensemble and pre-trained BERT text embeddings instead of DIME | Image-text retrieval · Microsoft COCO (5K) | Image retrieval R@1: 43.1 -> 43.6 | Helped |
| Add action-score predictions (ablation) | AME · Flickr30K | Text retrieval R@1: 73.8 -> 74.9 | Helped |
| Add action-score predictions (ablation) | AME · Flickr30K | Image retrieval R@1: 58.2 -> 58.9 | Helped |
| Add the action-aware memory bank (ablation) | AME · Flickr30K | Text retrieval R@1: 70.1 -> 74.9 | Helped |
| Add the action-aware memory bank (ablation) | AME · Flickr30K | Image retrieval R@1: 54.3 -> 58.9 | Helped |
| Add both action-score predictions and the action-aware memory bank (ablation) | AME · Flickr30K | Text retrieval R@1: 69.2 -> 74.9 | Helped |
| Add both action-score predictions and the action-aware memory bank (ablation) | AME · Flickr30K | Image retrieval R@1: 52.9 -> 58.9 | Helped |
| Increase the number of action-tag candidates (ablation) | AME · Flickr30K | Text retrieval R@1 | Mixed |
| Increase the number of action-tag candidates (ablation) | AME · Flickr30K | Image retrieval R@1 | Mixed |
| Increase the number of memory response texts (ablation) | AME · Flickr30K | Text retrieval R@1 | Helped |
Closest papers by meaning
- Text Is MASS: Modeling as Stochastic Embedding for Text-Video Retrieval (CVPR 2024 · similarity 0.84)
- Highlighting What Matters: Promptable Embeddings for Attribute-Focused Image Retrieval (NeurIPS 2025 · similarity 0.84)
- Exploiting Distribution Constraints for Scalable and Efficient Image Retrieval (ICLR 2025 · similarity 0.83)
- OST: Refining Text Knowledge with Optimal Spatio-Temporal Descriptor for General Video Recognition (CVPR 2024 · similarity 0.83)
- Align Before Adapt: Leveraging Entity-to-Region Alignments for Generalizable Video Action Recognition (CVPR 2024 · similarity 0.83)
- TIGeR: Unifying Text-to-Image Generation and Retrieval with Large Multimodal Models (ICLR 2025 · similarity 0.83)
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.