Research map / Vision-language and multimodal / Human and embodied modeling
DeepImageSearch: Benchmarking Multimodal Agents for Context-Aware Image Retrieval in Visual Histories
Chenlong Deng, Mengjie Deng, Junjie Wu, Dun Zeng, Teng Wang, Qingsong Xie, Jiadeng Huang, Shengjie Ma, Changwang Zhang, Zhaoxiang Wang, Jun Wang, Yutao Zhu and 1 more
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 GetMetadata tool (ablation) | ImageSeeker (Gemini-3-Flash-Preview) · DISBench | F1 (overall): 31.1 -> 36.8 | Improved |
| add FilterMetadata tool (ablation) | ImageSeeker (Gemini-3-Flash-Preview) · DISBench | F1 (overall): 31.8 -> 36.8 | Improved |
| add ViewPhotos tool (ablation) | ImageSeeker (Gemini-3-Flash-Preview) · DISBench | F1 (overall): 32.9 -> 36.8 | Improved |
| add WebSearch tool (ablation) | ImageSeeker (Gemini-3-Flash-Preview) · DISBench | F1 (overall): 32.2 -> 36.8 | Improved |
| add explicit state memory (ablation) | ImageSeeker (Gemini-3-Flash-Preview) · DISBench | F1 (overall): 31.9 -> 36.8 | Improved |
| add memory compression (ablation) | ImageSeeker (Gemini-3-Flash-Preview) · DISBench | F1 (overall): 33.9 -> 36.8 | Improved |
| add explicit state memory (ablation) | ImageSeeker (Gemini-3-Flash-Preview) · DISBench | F1 (inter-event): intra 34.1 -> 42.4; inter 30.1 -> 31.9 | Improved |
| increase parallel instances N with Best@k selection | ImageSeeker (Gemini-3-Flash-Preview) · DISBench | F1: 35.4 -> 60.8 | Improved |
| increase parallel instances N with Majority Voting | ImageSeeker (Gemini-3-Flash-Preview) · DISBench | F1 | Improved |
| switch embedding model from 2B to 8B | ImageSeeker agents · DISBench | F1 (overall): Claude-Sonnet-4.5 39.4 -> 43.8; GPT-5.2 35.1 -> 30.0 | Mixed |
Closest papers by meaning
- Navigating Massive Visual Context in Retrieval-Augmented Generation via Multimodal Memory Graph (ICML 2026 · similarity 0.90)
- ARK: A Dual-Axis Multimodal Retrieval Benchmark along Reasoning and Knowledge (NeurIPS 2026 · similarity 0.90)
- MMSearch-Plus: Benchmarking Provenance-Aware Search for Multimodal Browsing Agents (ICLR 2026 · similarity 0.89)
- OpenSearch-VL: An Open Recipe for Frontier Multimodal Search Agents (NeurIPS 2026 · similarity 0.89)
- SenseSearch: Empowering Vision-Language Models with High-Resolution Agentic Search-Reasoning via Reinforcement Learning (CVPR 2026 · similarity 0.89)
- RMIR: A Benchmark Dataset for Reasoning-Intensive Multimodal Image Retrieval (CVPR 2026 · similarity 0.89)
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.
