Jun 2026· arXiv.org· Vol abs/2606.22681· 3 citations· 47 references
Computer Science
TL;DR
GDP-RAG is proposed, a plan-based framework that targets only the information delta based on three simple design choices, and achieves the highest accuracy among all compared systems while maintaining a cost-of-pass of 0.51.
Abstract
Multi-hop question answering remains challenging for Retrieval-Augmented Generation (RAG) because existing approaches either propagate errors across iterative retrieval rounds or over-generate reasoning steps, increasing cost without improving accuracy. We propose Grounded Delta Planning RAG (GDP-RAG), a plan-based framework that targets only the information delta based on three simple design choices: (1) preliminary retrieval to ground planning before execution, (2) a gap-conditioned planning prompt that asks only for missing information, and (3) a skeletal trajectory that pairs each subquery with a Thought capturing evidence from preliminary retrieval and carrying it through to the final answer. GDP-RAG focuses computation on unresolved gaps, yielding concise, reliable reasoning trajectories. Extensive experiments on HotpotQA, 2WikiMultiHopQA, and MuSiQue show that GDP-RAG achieves the highest accuracy (60.63%) among all compared systems while maintaining a cost-of-pass of 0.51, 22% lower than PAR-RAG (0.65) and 68% lower than KnowTrace (1.57), with no method achieving both higher accuracy and lower cost.
Retrieval-Augmented Generation (RAG) effectively grounds large language models (LLMs) in external knowledge but struggles with
exploratory reasoning problems (ERPs)
that are the complex queries involving high uncertainty and ambiguity. Resolving ERPs requires complex reasoning with unclear paths, tending to result...
Gan-Lin Xu, Ling-Hao Zhang, Zhi-Tao Yin et al.· Proceedings of the ACM on Ma...· 0 citations
Retrieval-Augmented Generation (RAG) improves Large Language Models (LLMs) by incorporating external documents, but its effectiveness degrades when retrieved passages are irrelevant, incomplete, or inconsistent with parametric knowledge. We present LPDG-RAG, a post-training framework that extends reasoning-based knowle...
Hai-Feng Liu, Rui-Qi Zhao, Nan Zhao et al.· IEEE Transactions on Audio,...· 0 citations
Under this setup, learned planning is a weak relevance signal rather than a replacement for strong retrieval, and packing-order and score-flatness analyses did not identify a stable mechanism.
Ying-Rui Li, Han-Xiu Chen· Proceedings of the Seventh W...· 0 citations
Multi-hop question answering aims to answer complex queries by identifying and integrating evidence distributed across multiple documents. Existing iterative RAG approaches address this task by decomposing complex questions into plans or sub-questions, followed by repeated retrieval, evidence extraction, and reasoning...
Jumin Cho, Chanjun Park, Seung-Hoon Na et al.· Proceedings of the 35th ACM...· 0 citations
Trajectory-Aware Evidence Coordination (TAEC), a training-free framework that coordinates evidence use around unresolved answer requirements, achieves the best overall performance against leading training-free visual RAG baselines, with the highest average accuracy across multiple proprietary vision-language models.
Ya-Lun Wu, Bing-Zhou Wang, Bo-Yang Wang et al.· 1 citation
Question answering over public, historical open sources requires reliable retrieval behavior and access to exact structured records. We present SADA-RAG, a retrieval-augmented question answering (QA) framework for GEO-OSINT question answering over public historical sources that combines staged agent adaptation with det...