BMGQ: A Bottom-up Method for Generating Complex Multi-hop Reasoning Questions from Semi-structured Data
Bingsen Qiu, Zijian Liu, Xiao Liu, Bingjie Wang, Feier Zhang, Yixuan Qin, Chunyan Li, Haoshen Yang, Zeren Gao

TL;DR
BMGQ is an automated bottom-up approach that generates complex multi-hop questions from semi-structured data, enhancing training datasets for advanced question answering models with minimal human effort.
Contribution
The paper introduces BMGQ, a scalable, automated method for creating challenging multi-hop questions from semi-structured sources, addressing data bottlenecks in training retrieval and reasoning models.
Findings
Produces high-difficulty, verifiable questions suitable for training and evaluation.
Reduces human curation effort significantly.
Maintains the complexity and challenge of evaluation benchmarks.
Abstract
Building training-ready multi-hop question answering (QA) datasets that truly stress a model's retrieval and reasoning abilities remains highly challenging recently. While there have been a few recent evaluation datasets that capture the characteristics of hard-to-search but easy-to-verify problems -- requiring the integration of ambiguous, indirect, and cross-domain cues -- these data resources remain scarce and are mostly designed for evaluation, making them unsuitable for supervised fine-tuning (SFT) or reinforcement learning (RL). Meanwhile, manually curating non-trivially retrievable questions -- where answers cannot be found through a single direct query but instead require multi-hop reasoning over oblique and loosely connected evidence -- incurs prohibitive human costs and fails to scale, creating a critical data bottleneck for training high-capability retrieval-and-reasoning…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Code & Models
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
