iCog Labs

help edit space_dashboard
Draft — This content has not been approved for publication.

← Back to Ecosystem

GitHub: github.com/iCog-Labs-Dev

Status: Active. Ethiopian AI research lab. Major contract R&D partner to SingularityNET, contributing 49+ repositories across MOSES, ECAN, bioinformatics, search, Galaxy platform, and MORK SDKs.

iCog Labs is an Addis Ababa-based artificial intelligence research and development company and one of the largest external contributors to the Hyperon ecosystem. Operating as a contract R&D partner to SingularityNET, iCog Labs develops MeTTa-native reimplementations of core algorithms, developer tooling, bioinformatics pipelines, and enterprise integrations.

Core Contributions

  • Evolutionary program learning: metta-moses — full MeTTa reimplementation of MOSES (Meta-Optimizing Semantic Evolutionary Search); MOSES-MORK — MOSES integration with MORK backend; moses-optimization-algorithms — C++ optimization algorithm library
  • Attention allocation: metta-attention — MeTTa port of Economic Attention Networks (ECAN) with mind agents
  • Pattern mining: hyperon-miner extensions and pm-tutorial — interactive pattern miner tutorial
  • MORK SDKs: mork-rust-sdk (Rust) and mork-ts-sdk (TypeScript) — client SDKs for the MORK API

Developer Tooling

  • MeTTa LSP: metta-lsp — VS Code language server with Tree-sitter parsing, diagnostics, and navigation
  • Training portal: mettatraining — 51-lesson MeTTa curriculum with progressive challenges
  • MeTTa AI Assistant: RAG-based coding assistant (FastAPI + React) for MeTTa programming
  • Utility libraries: meTTa-utils and petta-utils — data structures and helper functions for MeTTa and PeTTa

Bioinformatics and Knowledge Graphs

  • BioChatter-MeTTa: NL-to-MeTTa query converter for biomedical knowledge graphs (client/server architecture)
  • Bio-semantic parsing: LLM-based pipelines for extracting biological datasets (DrugAge, GEO) into MeTTa and FOL
  • PubChem2MeTTa: PubChem RDF to MeTTa via BioCypher
  • Galaxy platform: Python library, Docker deployment, and Gemini-powered tool/workflow recommender for Galaxy bioinformatics servers

Search and Enterprise

  • Semantic search: Python semantic search API, Mattermost plugins (Go + React), and Django RAG API for document querying
  • Cognitive architectures: PeTTa-OpenPSI — PeTTa port of OpenPsi; MetaMo-Python — category theory motivational framework

Research

  • Predictive coding: PC-Transformers and NGC-PC-Transformers — local-learning transformer architectures
  • Quantum computing: Hybrid quantum-classical ML (PennyLane + PyTorch); FCA + QUBO optimization
  • Blockchain: DAOMatcher (DAO participant matching); Plutus stablecoin smart contracts

Related cards: SingularityNET · MOSES · ECAN · MORK · Bioinformatics

Key References



Discussion