Joaqu\'in Polonuer (Department of Biomedical Informatics, Harvard Medical School, Boston, Ma, USA, Departamento de Computaci\'on, FCEyN, Universidad de Buenos Aires, Buenos Aires, Argentina), Lucas Vittor (Department of Biomedical Informatics, Harvard Medical School, Boston, Ma, USA), I\~naki Arango (Department of Biomedical Informatics, Harvard Medical School, Boston, Ma, USA), Ayush Noori (Department of Biomedical Informatics, Harvard Medical School, Boston, Ma, USA, Department of Engineering Science, University of Oxford, Oxford, UK), David A. Clifton (Department of Engineering Science, University of Oxford, Oxford, UK, Oxford Suzhou Centre for Advanced Research, University of Oxford, Suzhou, Jiangsu, China), Luciano Del Corro (ELIAS Lab, Departamento de Ingenier\'ia, Universidad de San Andr\'es, Victoria, Argentina, Lumina Labs, Buenos Aires, Argentina), Marinka Zitnik (Department of Biomedical Informatics, Harvard Medical School, Boston, Ma, USA, Kempner Institute for the Study of Natural and Artificial Intelligence, Allston, Ma, USA, Broad Institute of MIT and Harvard, Cambridge, Ma, USA, Harvard Data Science Initiative, Cambridge, Ma, USA)

Retrieving evidence for language model queries from knowledge graphs requires balancing broad search across the graph with multi-hop traversal to follow relational links. Similarity-based retrievers provide coverage but remain shallow, whereas traversal-based methods rely on selecting seed nodes to start exploration, which can fail when queries span multiple entities and relations. We introduce ARK: Adaptive Retriever of Knowledge, an agentic KG retriever that gives a language model control over this breadth-depth tradeoff using a two-operation toolset: global lexical search over node descriptors and one-hop neighborhood exploration that composes into multi-hop traversal. ARK alternates between breadth-oriented discovery and depth-oriented expansion without depending on a fragile seed selection, a pre-set hop depth, or requiring retrieval training. ARK adapts tool use to queries, using global search for language-heavy queries and neighborhood exploration for relation-heavy queries. On STaRK, ARK reaches 59.1% average Hit@1 and 67.4 average MRR, improving average Hit@1 by up to 31.4% and average MRR by up to 28.0% over retrieval-based and agentic training-free methods. Finally, we distill ARK's tool-use trajectories from a large teacher into an 8B model via label-free imitation, improving Hit@1 by +7.0, +26.6, and +13.5 absolute points over the base 8B model on AMAZON, MAG, and PRIME datasets, respectively, while retaining up to 98.5% of the teacher's Hit@1 rate.

Sign up or login to leave a comment