Scientific discoveries are often framed as filling gaps between existing ideas. A new preprint on arXiv proposes a way to find such gaps systematically by treating knowledge as a geometric structure. The authors embed scientific concepts in a high-dimensional space and apply persistent homology, a tool from topological data analysis, to detect holes—regions where connections between ideas are missing.
The paper's title highlights a striking pattern: while filling simple, low-order holes tends to draw collective attention, most higher-order holes remain unexplored. These are more complex gaps that may represent opportunities for deeper or more unusual connections. The method offers a way to map the landscape of unknown science, potentially guiding researchers toward neglected areas.
As a preprint, the work has not yet undergone peer review, and the abstract provides limited detail on the data or validation. Still, the approach suggests a novel way to quantify the structure of scientific ignorance, not just its discoveries.