DeepLog: the PyTorch for neurosymbolic AI
A state of the art Weighted and Projected Sampler
A Python (re-)implementation of some known knowledge compilers
A Knowledge Compilation-based Uniform Sampler
Maintain a centralized knowledge wiki for LLMs by compiling raw documentation into actionable, versioned markdown files.
Python code accompanying my thesis on static vtree heuristics for SDD compilation.
CompiledKnowledge is a Python package for compiling and querying discrete probabilistic graphical models.
An offline-first, model-agnostic local inbox ingestion pipeline and compiler for Obsidian vaults, inspired by Karpathy's LLM Wiki paradigm.
Compile system models to d-DNNF circuits, then diagnose, track, and learn on one artifact: exact posteriors, ranked diagnoses, mode tracking, EM — with a PyTorch backend for GPU inference.
Paper + experiments: compiled-wiki lifecycle with 4 stages (compile, certify, maintain, retract) for LLM-compiled wikis, providing certified correctness guarantees over evolving corpora. Includes minimal llm-wiki implementation.
To associate your repository with the knowledge-compilation topic, visit your repo's landing page and select "manage topics."