Formal Concept Lattices as Semantic Scaffolds for Concept-Based Learning

Arxiv CS.CV (computer vision)

Paper proposes using Formal Concept Lattices (FCLs) as interpretable semantic scaffolds for concept-based learning, achieving improved alignment with human reasoning. Experiments show FCL-based concept representations outperform standard methods on multiple benchmarks.

Arxiv CS.CV (computer vision)·Deepika SN VemuriSayanta AdhikariAnkit SahaKrishn Vishwas KherVineeth N Balasubramanian

6 days ago