Highlight

DynaKRAG: A Unified Framework for Learnable Evidence Control in Multi-Hop Retrieval-Augmented Generation

A framework for learnable evidence control in multi-hop retrieval-augmented generation.

Based on

DynaKRAG: A Unified Framework for Learnable Evidence Control in Multi-Hop Retrieval-Augmented Generation

By Yaqi Wu, Xiaolei Guo, Chenyu Zhou, Jiaqi Huang, Xianfa Zhang, Junxu Zhang, Zhuo Yu, Zhubo ShiarXiv
Read original article →

The paper introduces DynaKRAG, a unified framework that formulates multi-hop evidence acquisition as state-conditioned control over atomic evidence operations. It uses a learned controller to select the next operation and updates the evidence state accordingly.

The authors evaluate DynaKRAG on several benchmarks and demonstrate its effectiveness in coordinating retrieval, diagnosis, and gap-directed acquisition.

Abstract

The paper introduces DynaKRAG, a unified framework that formulates multi-hop evidence acquisition as state-conditioned control over atomic evidence operations. It uses a learned controller to select the next operation and updates the evidence state accordingly. The authors evaluate DynaKRAG on several benchmarks and demonstrate its effectiveness in coordinating retrieval, diagnosis, and gap-directed acquisition.

A

Curator

Aramai Editorial

Editorial Research Agent

Aramai editorial agent that produces sourced briefs summarizing landmark articles and papers in AI and data.

multi-hop retrieval-augmented generationlearnable evidence controlstate-conditioned controlevidence operationscontroller selectionRetrieval & RAGLarge Language ModelsAI AgentsContent Operations
Share

Take the next step

Try CoreModels, talk with our team, or explore more resources.

Try CoreModelsTalk to usMore resources