Research map / Vision-language and multimodal / Human and embodied modeling
Visual Grounding in Video for Unsupervised Word Translation
Gunnar A. Sigurdsson, Jean-Baptiste Alayrac, Aida Nematzadeh, Lucas Smaira, Mateusz Malinowski, Joao Carreira, Phil Blunsom, Andrew Zisserman
Check 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 |
|---|---|---|---|
| Use MUVE instead of VecMap for unsupervised word mapping | English-French word translation · English-French Dictionary, All | Recall@1: 28.4 -> 28.9 | Improved |
| Use MUVE instead of VecMap for unsupervised word mapping | English-Korean word translation · English-Korean Dictionary | Recall@1: 13.0 -> 17.7 | Improved |
| Use MUVE instead of VecMap for unsupervised word mapping | English-Japanese word translation · English-Japanese Dictionary | Recall@1: 13.7 -> 15.1 | Improved |
| Use the shared visual-grounding Base Model instead of video retrieval to create word mappings | English-French word translation · English-French Dictionary, All | Recall@1: 6.3 -> 9.1 | Improved |
| Use the shared visual-grounding Base Model instead of video retrieval to create word mappings | English-French word translation · English-French Simple Words, All | Recall@1: 12.5 -> 28.0 | Improved |
| Add the MUVE visual initialization to text-based word mapping | English-French Dictionary, All · English-French Dictionary | Recall@1: 28.4 -> 28.9 | Improved |
| Add the MUVE visual initialization to text-based word mapping | English-French Simple Words, All · English-French Simple Words | Recall@1: 28.0 -> 58.3 | Improved |
| Add the MUVE visual initialization to text-based word mapping | English-French Simple Words, Visual · English-French Simple Words, Visual | Recall@1: 45.3 -> 67.5 | Improved |
| Add the orthogonal constraint to the AdaptLayer (ablation) | MUVE · English-French Dictionary | Recall@1: 16.6 -> 28.9 | Improved |
| Use visual grounding instead of text-only alignment when English and French training corpora differ | MUVE versus text-based methods · English-French Dictionary, Wiki-En and HowToW-Fr | Recall@10: 32.6; text-based methods less than 0.5 | Improved |
| Reduce the amount of French training data | MUVE versus MUSE and VecMap · English-French Dictionary, reduced training data | Recall@10 | Improved |
| Reduce the French vocabulary size | MUVE versus MUSE and VecMap · English-French Dictionary, reduced vocabulary | Recall@10 | Improved |
Closest papers by meaning
- HERO: Hierarchical Embedding-Refinement for Open-Vocabulary Temporal Sentence Grounding in Videos (CVPR 2026 · similarity 0.89)
- MC-Bench: A Benchmark for Multi-Context Visual Grounding in the Era of MLLMs (ICCV 2025 · similarity 0.88)
- EGM: Efficient Visual Grounding Language Models (ECCV 2026 · similarity 0.88)
- VideoGrounding-DINO: Towards Open-Vocabulary Spatio-Temporal Video Grounding (CVPR 2024 · similarity 0.88)
- GroundingME: Exposing the Visual Grounding Gap in MLLMs through Multi-Dimensional Evaluation (CVPR 2026 · similarity 0.88)
- Empower Words: DualGround for Structured Phrase and Sentence-Level Temporal Grounding (NeurIPS 2025 · 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.
