Search papers, labs, and topics across Lattice.
This paper introduces KBevo, a co-evolving framework that simultaneously constructs a structured knowledge base and enhances reasoning capabilities for knowledge-intensive question answering. By optimizing both the knowledge base and reasoning processes end-to-end using QA outcome rewards, KBevo significantly improves the quality and connectivity of the knowledge structures. The results show that this approach not only increases answer reachability but also enhances compositional factual reasoning and controllability compared to traditional retrieval methods.
Co-evolving knowledge bases with reasoning capabilities can dramatically enhance the accuracy and relevance of answers in knowledge-intensive tasks.
Retrieval-augmented methods improve factual accuracy by grounding language models in external knowledge, but retrieving over unstructured text often introduces irrelevant context and offers limited control over the retrieved information. Structured knowledge bases offer a more controllable alternative, yet they are expensive to construct and often brittle to reason over. To address these limitations, we propose KBevo: a co-evolving framework that jointly learns to construct a structured knowledge base and reason over it for knowledge-intensive question answering. By optimizing both components end-to-end with QA outcome rewards, our method enables reasoning success to directly improve the quality of the constructed knowledge base. This leads to larger, better-connected knowledge structures with higher answer reachability, while also improving compositional factual reasoning and controllability compared to standard retrieval baselines.
