Scientific discovery guided by multiobjective search in knowledge space
Knowledge Manifold Optimization for Scientific Discovery
Information RetrievalMachine Learning
Summary
Scientific literature is growing fast, making it hard to see what's known and where new discoveries are possible. The authors created a method called Knowledge Manifold Optimization that looks for promising discovery spots by considering many factors like how well-supported an area is and its novelty. Instead of just picking one best spot, their method finds a balanced set of good spots for new research. They tested it on aerodynamics papers and turned some findings into clear research ideas using a language model.
What this means in practice
- •For research software developers: Create tools that highlight promising, well-supported areas for new scientific research using multiobjective search instead of single-score rankings.
- •For scientific publishing platforms: Integrate evidence-based discovery recommendations to guide authors toward unexplored yet supported research topics in specialized literature.
Tested on one dataset.
Authors
Shigeru Obayashi, Yoshiaki Kawagoe, Kazuhiko Komatsu, Tomonaga Okabe
Abstract
The rapid expansion and specialization of scientific literature make it increasingly difficult to identify not only what is known, but where new and still supportable knowledge may emerge. We introduce Knowledge Manifold Optimization (KMO), a discovery-oriented extension of the Knowledge Manifold framework that treats scientific discovery as a multiobjective search over a continuous, literature-derived knowledge space. Candidate locations are evaluated using four complementary objectives: persistent knowledge-regime boundary strength, band-pass white-space novelty, semanticgradient magnitude, and evidential support. Rather than collapsing these criteria into a single weighted score, KMO identifies Pareto-optimal Supported Knowledge Gaps, defined as corpus-relative unoccupied regions that remain sufficiently supported by surrounding literature. The method is demonstrated on a 99-paper corpus in aerodynamics and aerodynamic design. On a 21 x 21 grid, exact Pareto analysis produced 19 non-dominated points, of which 16 formed the practical front after excluding points with zero white-space score. These solutions exhibited distinct discovery modes, including methodology bridges, aero-structural topology/FSI integration, functional aerodynamic-device integration, and rear-wake/vehicle flow-control integration. A distance-maximization baseline preferentially selected extrapolative regions with substantially lower support and reliability, demonstrating why emptiness alone is insufficient. Representative Pareto locations were converted into auditable evidence packages and then verbalized by a constrained large language model into candidate research hypotheses. The results show that scientific discovery can be formulated as structured multi-objective search in an explicit knowledge space, while keeping novelty corpus-relative and separating numerical discovery from downstream language generation.