A research & simulation environment for the Adaptive Intelligence Circle (AIC) ecosystem. AIC-Lab provides a visual sandbox for simulating, testing, and analyzing components of the AIC system, including packet pipeline, preprocessing, anomaly detection, introspection, behavior engine, and rollback logic — visualized using Jupyter Notebook and Google Colab.

This repository is designed to run in parallel with AIC Core (C++23) and serve as an algorithm development lab, enabling you to quickly test, simulate complex behavior, and experiment with ideas before deploying them to production systems.

AIC-Lab was created to serve 5 main goals:

• simulation pipeline from sensor → preprocessing → protection → rollback
• random packet generation, simulated attacks, error data
• analysis of internal system behavior and state

• introspection logic testing
• modeling behavior, reactions, confidence scores
• observing event statistics and rollback mechanisms

Run the same logic:

• generate_packet.py
• generate_packet.cpp

→ Cross-validate to check consistency, logic, and performance.

Helps see the whole picture of AIC operations:

• timing diagram
• state transition
• magnitude of anomaly
• heatmap for event pipeline

Record:

• packet processing time
• rollback latency
• throughput preprocessing
• introspection speed From there, optimize Core C++.

AIC-Lab/
│
├── notebooks/
│ ├── 01_intro.ipynb
│ ├── 02_generate_packet_test.ipynb
│ ├── 03_behavior_simulation.ipynb
│ ├── 04_anomaly_injection.ipynb
│ ├── 05_rollback_simulation.ipynb
│ ├── 06_cross_validation_cpp_python.ipynb
│ ├── 07_remote_slm_plugin_test.ipynb
│ └── 08_introspection_engine.ipynb
│
├── cpp/
│ ├── CMakeLists.txt
│ ├── aic_core_minimal.cpp
│ ├── generate_packet.cpp
│ ├── simulate_behavior.cpp
│ └── packet.h
│
├── python/
│ ├── generate_packet.py
│ ├── behavior_engine_sim.py
│ ├── preprocessing_sim.py
│ ├── introspection_sim.py
│ └── utils.py
│
├── data/
│ ├── test_packets.json
│ ├── anomaly_cases.json
│ └── behavior_rules.json
│
├── colab_setup.sh
└── README.md

Install extensions:

• Python
• Jupyter
• CMake Tools
• C/C++

Run Notebook directly in VSCode.

Run any notebook → will automatically install:

!bash colab_setup.sh

Includes:

• g++
• cmake
• build C++ scripts
• Python environment

Example in Colab:

%cd cpp
!cmake .
!make
!./aic_core_minimal

🧩 Create packet:

• timestamp
• payload
• entropy
• magnitude
• actor ID
• anomaly level

Ensure Python ↔ C++ logic is consistent through cross-testing.

📘 Simulation:

• rules
• states
• actions
• adaptive responses
• warning levels

Notebook helps visualize the entire lifecycle of a behavior.

🔍 Test:

• analyzeThoughtProcess
• summarizeIntrospection
• memory snapshot
• reasoning tree
• self-evaluation score

🔥 Simulate the situation:

• overload
• anomaly
• misbehavior
• policy violation → Then observe the process rollback → stabilize → recover.

🧼 Simulation:

• data cleaner
• feature extractor
• entropy filters
• windowed statistics

Repo supports exporting:

• anomaly graph over time
• behavior graph
• sensor value histogram
• flow graph pipeline
• ​​checkpoint timeline of rollback

AIC-Lab does not replace AIC-Core.

It only serves as:

🧪 Research lab 📈 Simulation sandbox 🔍 Introspection & validation tool 🚦 Development acceleration test suite

All the actual C++ is still in AIC-Core.

• packet simulation
• anomaly injection
• cross test Python ↔ C++

• rule-based simulation
• visualize state machine

• inference modeling
• reasoning tree flow

• hybrid rollback
• consistency check
• log timeline

• end-to-end sensor → rollback
• unified interface
• visualization toolkit

Anyone who wants to:

• simulate behavior
• design new plugins
• write demonstration notebooks
• test anomaly

AIC-Lab is open.

**“Who can also see a system working through code.

But very few people get to see inside it: its thinking, behavior, adaptation, and self-protection mechanisms.”**

AIC-Lab was born to open that door.