Research map / Vision-language and multimodal / Human and embodied modeling
Hierarchical Abstract Tree for Cross-Document Retrieval Augmented Generation
Ziwen Zhao, Menglin Yang
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 |
|---|---|---|---|
| replace k-means-type tree (RAPTOR) with Psi-RAG abstract tree and agentic hybrid retrieval | Tree-RAG · Single/multi-hop QA average | F1 (%): 36.69 -> 62.77 | Improved |
| Psi-RAG versus HippoRAG 2 on cross-document QA | Psi-RAG · Single/multi-hop QA average | F1 (%): 55.41 -> 62.77 | Improved |
| add retrieval and answering agent (ablation) | Psi-RAG · multi-hop QA | F1 (%): average gain of ~20% | Improved |
| add R&A agent (ablation) | Psi-RAG · NQ | F1 (%): +1.89% | Improved |
| add sparse keyword retrieval (BM25) (ablation) | Psi-RAG · PopQA | F1 (%): +21.82% | Improved |
| add sparse keyword retrieval (BM25) (ablation) | Psi-RAG · 2Wiki | F1 (%): +41.21% | Improved |
| use Qwen3-Reranker-8B instead of RRF for fusion (ablation) | Psi-RAG · QA datasets | F1 (%) | No clear effect |
| add query reorganization (ablation) | Psi-RAG · MuSiQue | F1 (%): 45.82 -> 47.83 | Improved |
| add query reorganization (ablation) | Psi-RAG · MultiHop-RAG | F1 (%): 55.03 -> 55.18 | No clear effect |
| use summative instead of keyword abstracts (ablation) | Psi-RAG · Single/multi-hop QA average | F1 (%): 62.38 -> 62.77 | No clear effect |
| replace Llama-3.3-70B abstraction agent with Llama-3.2-1B (ablation) | Psi-RAG · HotpotQA | F1 (%): 74.85 -> 73.23 | No clear effect |
| replace Llama-3.3-70B R&A agent with Gemma-3-27B (ablation) | Psi-RAG · 2Wiki and MuSiQue | F1 (%): 2Wiki 76.94 -> 71.78; MuSiQue 47.83 -> 49.30 | Mixed |
Closest papers by meaning
- Cooperative Retrieval-Augmented Generation for Question Answering: Mutual Information Exchange and Ranking by Contrasting Layers (NeurIPS 2025 · similarity 0.90)
- RAPTOR: Recursive Abstractive Processing for Tree-Organized Retrieval (ICLR 2024 · similarity 0.90)
- Cross-Granularity Hypergraph Retrieval-Augmented Generation for Multi-hop Question Answering (AAAI 2026 · similarity 0.90)
- LP-RAG: Learning to Retrieve with Link Predictors (NeurIPS 2026 · similarity 0.89)
- GFM-RAG: Graph Foundation Model for Retrieval Augmented Generation (NeurIPS 2025 · similarity 0.88)
- UDA: A Benchmark Suite for Retrieval Augmented Generation in Real-World Document Analysis (NeurIPS 2024 · 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.
