Research

Research in AGI architecture
and cognitive systems.

iCog Labs studies the technical layers required for advanced AI: memory, attention, motivation, symbolic discovery, program learning, reasoning, verified knowledge ingestion, adaptive learning, and reusable skill safety.

Anchor
Neural-symbolic AGI
Method
Neural-symbolic systems
Purpose
Safe, beneficial AGI

Core technical areas in the lab's AGI research program.

  1. 01
    Program Evolution

    MOSES

    An AI research project exploring how machines can discover programs and strategies that solve problems with less human programming.

    Explore project
    Program synthesis / MeTTa
  2. 02
    Motivational Architecture

    OpenPsi

    An implementation of Dietrich Dorner's PSI theory in OpenCog Hyperon, connecting motivation, emotion, and decision-making.

    Explore project
    Motivation / emotion / behavior
  3. 03
    Symbolic discovery

    Hyperon Pattern Miner

    A core cognitive component that discovers frequent and surprising structures in Atomspace, turning implicit graph activity into reusable symbolic knowledge.

    Explore project
    Pattern mining / habit formation
  4. 04
    Quantum machine learning

    Quantum ML

    Research spanning quantum machine learning for drug discovery and hardware-aware circuit synthesis using ZX-calculus and Hyperon frameworks.

    Quantum AI / circuit synthesis
  5. 05
    Adaptive learning

    Continual Learning

    An inference-oriented approach to cumulative learning using sequential Bayesian inference, predictive coding, uncertainty, and causal modularity.

    Explore project
    Bayesian inference / predictive coding

Research translated into deployed AI systems.

The products page shows how this research informs platforms for biology, data, media, reasoning, and generative systems.

View products