How jiaxiaogang/he works
he4o is an iOS-oriented embodied-AI system intended to learn and act in a changing environment. Its documented loop takes sensory input, recognizes and learns network-based representations, manages needs and plans, chooses and emits behavior, then uses feedback for further learning. The graph preserves major documented stages and vision capability. Most implementation files were not sampled, so documented README relationships are cited there; the graph does not infer calls between unsampled modules.
Main entry points: AINet.h Knowledge network · TCRecognition.h Recognition · AIInput.h Sensory input · AIVisionAlgs.h Vision algorithms
Components 16
Experience
- Sensory input
SMG_NothingIsAll/AIFoundation/AIInput/AIInput.h - Vision algorithms
SMG_NothingIsAll/CortexAlgorithms/Vision/AIVisionAlgs.h - Recognition
SMG_NothingIsAll/AIFoundation/AIThinkingControl/TCRecognition/TCRecognition.h - Short-term match
SMG_NothingIsAll/AIFoundation/AIThinkingControl/In/Model/ShortMatchManager.h
Knowledge Network
- Knowledge network
SMG_NothingIsAll/AIFoundation/AINet/AINet.h - Feature nodes
SMG_NothingIsAll/AIFoundation/AINet/AINode/AIFeatureNode.h - Abstraction builders
SMG_NothingIsAll/AIFoundation/AINet/AIAbsFoManager/AIAbsFoManager.h - Concept builders
SMG_NothingIsAll/AIFoundation/AINet/AIAlgNodeManager/AIAlgNodeManager.h
Thinking Control
- Demand manager
SMG_NothingIsAll/AIFoundation/AIThinkingControl/TCDemand/Model/DemandManager.h - Planning
SMG_NothingIsAll/AIFoundation/AIThinkingControl/TCPlan/TCPlan.h - Problem solving
SMG_NothingIsAll/AIFoundation/AIThinkingControl/TCSolution/TCSolution.h - Decision control
SMG_NothingIsAll/AIFoundation/AIThinkingControl/AIThinkingControl.h
Action and Learning
- Behavior output
SMG_NothingIsAll/AIFoundation/AIOutput/Output.h - Feedback
SMG_NothingIsAll/AIFoundation/AIThinkingControl/TCFeedback/TCFeedback.h - Learning and transfer
SMG_NothingIsAll/AIFoundation/AIThinkingControl/TCLearning/TCLearning.h
Other components
- World / user
Connections 18
- World / user → Sensory input: provides input
README.md - Sensory input → Recognition: feeds perception
README.md - Sensory input → Vision algorithms: supports vision
README.md - Vision algorithms → Recognition: supports recognition
README.md - Recognition → Short-term match: matches input
README.md - Recognition → Knowledge network: forms representations
README.md - Knowledge network → Feature nodes: represents features
README.md - Knowledge network → Abstraction builders: builds abstractions
README.md - Knowledge network → Concept builders: builds concepts
README.md - Short-term match → Demand manager: informs tasks
README.md - Demand manager → Planning: supports planning
README.md - Planning → Problem solving: seeks solutions
README.md - Problem solving → Decision control: supports decisions
README.md - Decision control → Behavior output: selects behavior
README.md - Behavior output → World / user: acts on world
README.md - World / user → Feedback: provides outcomes
README.md - Feedback → Learning and transfer: guides learning
README.md - Learning and transfer → Knowledge network: evolves knowledge
README.md