This repository documents a structured study of informed machine learning through reading, note-taking, and small-scale reproductions. It serves as a research workspace for organizing paper PDFs, bilingual note files, and toy implementations under a stable directory structure. The emphasis is on academic traceability rather than interface design or presentation. Core materials are grouped by topic so that papers, notes, and code references can be updated incrementally over time. The current focus is on survey reading, logic-guided learning, semantic constraints, and declarative learning systems.