Research map / Large language models / AI and society
Towards Personalized Deep Research: Benchmarks and Evaluations
Yuan Liang, Jiaxian Li, Yuqing Wang, WANG PIAOHONG, Motong Tian, Pai Liu, Shuofei Qiao, Runnan Fang, He Zhu, Ge Zhang, Minghao Liu, Yuchen Eleanor Jiang and 2 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 |
|---|---|---|---|
| provide user context instead of task only (ablation) | OAgents · PDR-Bench personalization | P-Score: 6.17 -> 6.53 | Improved |
| provide explicit persona instead of context (ablation) | OAgents · PDR-Bench personalization | P-Score: 6.53 -> 6.78 | Improved |
| provide explicit persona instead of context (ablation) | O3 Deep Research · PDR-Bench personalization | P-Score: 5.48 -> 5.46 | No clear effect |
| provide explicit persona instead of task only (ablation) | Gemini-2.5-Pro w/Search · PDR-Bench personalization | P-Score: 3.96 -> 4.70 | Improved |
| add O-Mem memory system to infer persona from context (ablation) | Perplexity Deep Research · PDR-Bench Task w/Context | P-Score: 3.69 -> 4.26 | Improved |
| add Mem0 memory system (ablation) | Perplexity Deep Research · PDR-Bench Task w/Context | P-Score: 3.69 -> 3.55 | Hurt |
| add Memory OS memory system (ablation) | Perplexity Deep Research · PDR-Bench Task w/Context | P-Score: 3.69 -> 3.88 | Improved |
| equip an LLM with search instead of a dedicated deep research agent | LLMs with search · PDR-Bench Task w/Persona | Overall score: 6.64 -> 5.53 (best OAgents vs best search LLM) | Hurt |
Closest papers by meaning
- MiroEval: Benchmarking Multimodal Deep Research Agents in Process and Outcome (NeurIPS 2026 · similarity 0.86)
- DRBench: A Realistic Benchmark for Enterprise Deep Research (ICLR 2026 · similarity 0.86)
- MMPB: It’s Time for Multi-Modal Personalization (NeurIPS 2025 · similarity 0.85)
- MLR-Bench: Evaluating AI Agents on Open-Ended Machine Learning Research (NeurIPS 2025 · similarity 0.85)
- From Intent to Evidence: A Categorical Approach for Structural Evaluation of Deep Research Agents (NeurIPS 2026 · similarity 0.85)
- HYDRA: Model Factorization Framework for Black-Box LLM Personalization (NeurIPS 2024 · similarity 0.85)
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.
