Learning to Construct Knowledge through Sparse Reference Selection with Reinforcement Learning

2025-09-06

AI Summary

The exponential growth of scientific literature impedes domain-specific knowledge acquisition due to complex reasoning requirements, limited access to full texts, and extreme sparsity of target references within vast candidate pools. To address this, we propose a deep reinforcement learning (DRL)-based approach for sparse citation selection that emulates human-like, incremental knowledge construction under information constraints—relying solely on local textual cues such as titles and abstracts to efficiently identify the most informative papers. This work represents the first systematic application of DRL to sparse citation selection in scientific literature. Evaluated on drug–gene relationship discovery, our method significantly improves retrieval precision and accelerates knowledge discovery, demonstrating that both the model and humans can effectively build structured knowledge from restricted textual inputs. The approach establishes a novel paradigm for low-resource scientific intelligence mining.

Technology Category

• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Knowledge Representation and Reasoning: Knowledge Acquisition
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies

Abstract

The rapid expansion of scientific literature makes it increasingly difficult to acquire new knowledge, particularly in specialized domains where reasoning is complex, full-text access is restricted, and target references are sparse among a large set of candidates. We present a Deep Reinforcement Learning framework for sparse reference selection that emulates human knowledge construction, prioritizing which papers to read under limited time and cost. Evaluated on drug--gene relation discovery with access restricted to titles and abstracts, our approach demonstrates that both humans and machines can construct knowledge effectively from partial information.

Problem

• Selecting sparse references from vast literature
• Prioritizing papers under limited time constraints
• Constructing knowledge from restricted text access

Innovation

• Deep Reinforcement Learning for reference selection
• Prioritizes papers to read under constraints
• Effective knowledge construction from partial information