Research map / Vision-language and multimodal / Cross-modal learning
Guided Query Refinement: Multimodal Hybrid Retrieval with Test-Time Optimization
Omri Uzan, Asaf Yehudai, roi pony, Eyal Shnarch, Ariel Gera
Read the paperCheck 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 |
|---|---|---|---|
| add GQR with Jina text complementary retriever | Colnomic-7B · ViDoRe 2 | NDCG@5: 60.3 -> 63.1 | Improved |
| add GQR with Linq-Embed complementary retriever | Llama-Nemo · ViDoRe 2 | NDCG@5: 63.0 -> 65.2 | Improved |
| add GQR with much weaker Qwen3 text retriever | Llama-Nemo · ViDoRe 2 | NDCG@5: 63.0 -> 63.3 | Improved |
| add GQR with Jina text complementary retriever | Colnomic-7B · ViDoRe 3 (zero-shot) | NDCG@5: 55.7 -> 57.4 | Improved |
| add GQR with Linq-Embed complementary retriever | Colnomic-7B · ViDoRe 1 | NDCG@5: 89.8 -> 89.8 | No clear effect |
| use GQR instead of rank-level fusion (RRF) | hybrid retrieval · ViDoRe 2 | NDCG gain over primary (%): -2.8 -> +3.9 | Improved |
| use GQR instead of tuned Min-Max score aggregation | hybrid retrieval · ViDoRe 2 | NDCG gain over primary (%): +3.4 -> +3.9 | Improved |
| tune fusion weight on dev set | Score Aggregation (Min-Max) · ViDoRe 2 | NDCG gain over primary (%): +0.4 -> +3.4 | Improved |
| run an extra full-index search with the refined query (ablation) | GQR · ViDoRe 2 | NDCG@5 | No clear effect |
| replace KL(p_avg||p_1) objective with JS or KL(p_2||p_1) (ablation) | GQR · ViDoRe 2 | NDCG@5 | No clear effect |
| restrict candidate pool to primary top-K instead of union (ablation) | GQR · ViDoRe 2 | NDCG@5 | No clear effect |
| use very large learning rate 5e-3 (ablation) | GQR · ViDoRe 2 | NDCG@5 | Hurt |
Closest papers by meaning
- Your Embedding Model Is SMARTer Than You Think (NeurIPS 2026 · similarity 0.89)
- MetaEmbed: Scaling Multimodal Retrieval at Test-Time with Flexible Late Interaction (ICLR 2026 · similarity 0.89)
- Document Haystacks: Vision-Language Reasoning Over Piles of 1000+ Documents (CVPR 2025 · similarity 0.89)
- MRAG-Bench: Vision-Centric Evaluation for Retrieval-Augmented Multimodal Models (ICLR 2025 · similarity 0.89)
- MRMR: A Realistic and Expert-Level Multidisciplinary Benchmark for Reasoning-Intensive Multimodal Retrieval (ICLR 2026 · similarity 0.88)
- RMIR: A Benchmark Dataset for Reasoning-Intensive Multimodal Image Retrieval (CVPR 2026 · similarity 0.88)
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.
