Axiocortex Research

Hierarchical World Models, Sample-Efficient Learning, and Brain-Inspired AI

Papers

Core references & recent advances

People

Key researchers to follow

Groups

Labs, universities & startups

News

Latest developments

About This Research

This hub tracks research on HICA (Hierarchical Intrinsic Curiosity Agent) — a brain-inspired approach to building sample-efficient AI agents. The core idea: combine hierarchical world models (like Dreamer), complementary learning systems (hippocampus + neocortex), and predictive maps (successor representation) to create agents that learn fast, generalize well, and plan hierarchically.

World Models Dreamer CLS Theory Successor Representation Hierarchical RL Sample Efficiency Predictive Coding

Paper Relationship Graph

World Models CLS HRL Spatial Predictive Robot Exploration

Click nodes for details - Drag to move - Scroll to zoom