Learning to Construct Knowledge through Sparse Reference Selection with Reinforcement Learning

This addresses the challenge of knowledge acquisition from rapidly expanding scientific literature for researchers in specialized domains, though it appears incremental as it applies reinforcement learning to an existing bottleneck.

The paper tackles the problem of acquiring new knowledge from scientific literature in specialized domains with complex reasoning, restricted full-text access, and sparse target references by presenting a Deep Reinforcement Learning framework for sparse reference selection that prioritizes which papers to read under limited time and cost. Evaluated on drug-gene relation discovery with access restricted to titles and abstracts, the approach demonstrates that both humans and machines can construct knowledge effectively from partial information.