An interactive educational simulation of the AION (Attunement through Iterative Oscillatory Networks) model — a quantum-inspired neural architecture where meaning emerges through phase synchronization among coupled oscillators.
This project provides a Streamlit-based web application that demonstrates the AION model through:
- Real-time visualization of oscillators on a unit circle
- Resonance index tracking over time
- Interactive parameter controls
- Educational content explaining key concepts
Core Concept: In the AION model, phase coupling represents how meaning emerges through interactions—oscillators "communicate" through their phase relationships, leading to synchronized behavior. This mirrors how neurons in the brain communicate through synchronized firing patterns to create coherent perceptions and meaning.
- Real-time Simulation: Watch oscillators synchronize through the Kuramoto model
- Interactive Controls: Adjust number of oscillators, coupling strength, time step, and animation speed
- Visualizations:
- Unit circle plot showing oscillator positions
- Resonance index chart tracking synchronization over time
- Educational Content: Learn about oscillators, phase, resonance, and attunement
- Python 3.11 or higher
- pip (Python package manager)
-
Clone the repository:
git clone <repository-url> cd Attunement-through-Iterative-Oscillatory-Networks
-
Install dependencies:
pip install -r requirements.txt
Start the Streamlit application:
streamlit run src/app.pyThe application will open in your default web browser at http://localhost:8501.
- Start Simulation: The simulation auto-starts when you load the page
- Observe Behavior: Watch oscillators converge and cluster together on the unit circle as they synchronize
- Monitor Resonance: Track the resonance index increasing from low values (0.1-0.3) toward 1.0 as synchronization occurs
- Control Simulation: Use pause/resume/reset buttons to control execution
Initial State:
- Oscillators start randomly distributed around the unit circle
- Resonance index R is typically low (0.1-0.3)
- Each oscillator rotates at its own intrinsic frequency
With Coupling (K > 0):
- Oscillators gradually converge and cluster together on the circle
- Resonance index R increases over time toward 1.0
- The cluster rotates around the circle as a synchronized group
- Higher coupling strength (K) leads to faster convergence
Perfect Synchronization:
- All oscillators cluster at approximately the same phase
- Resonance index R approaches 1.0
- The entire group rotates together as one unit
No Coupling (K = 0):
- Oscillators rotate independently
- They remain spread out around the circle
- Resonance index R stays low
- Number of Oscillators (N): Range [1, 1000] - Controls how many oscillators are in the network
- Coupling Strength (K): Range [0, 10] - Controls how strongly oscillators influence each other
- K = 0: No synchronization, oscillators rotate independently
- K = 2-3: Moderate coupling, gradual convergence and clustering
- K > 5: Strong coupling, rapid synchronization into tight cluster
- Time Step (Δt): Range (0, 1] - Controls simulation time increment
- Animation Speed: Range [0.1, 10.0] - Controls visual update rate
The application includes explanations for:
- Oscillator: A node in the network with a phase and intrinsic frequency. Watch how oscillators converge and cluster together as they synchronize.
- Phase: Angular position on the unit circle [0, 2π). As oscillators synchronize, their phases converge, causing them to cluster on the circle.
- Resonance: Degree of phase synchronization (0 = desynchronized/spread out, 1 = perfectly synchronized/clustered). The resonance index increases as oscillators converge.
- Attunement: The process of oscillators aligning their phases through coupling, resulting in the visible clustering behavior you observe in the visualization.
- To see convergence: Set coupling strength (K) to 2.0-4.0 and let the simulation run for 10-20 seconds
- To see faster clustering: Increase K to 5.0-7.0
- To see independent rotation: Set K to 0
- Watch the resonance chart: The line should trend upward as oscillators converge
- Check the unit circle: Blue dots should gradually cluster together over time
src/
├── aion/
│ ├── models/ # Core simulation logic
│ ├── visualization/ # Plotly visualization components
│ └── ui/ # Streamlit UI components
├── app.py # Main Streamlit entry point
└── config.py # Default configuration values
tests/
├── unit/ # Unit tests for core logic
└── integration/ # Integration tests for UI flows
Run the test suite:
pytestRun with coverage:
pytest --cov=src --cov-report=htmlThe phase update for oscillator j:
dθ_j/dt = ω_j + (K/N) Σ sin(θ_k - θ_j)
The coupling term (K/N) Σ sin(θ_k - θ_j) represents phase coupling—the mechanism by which
oscillators influence each other's phases. This phase coupling enables oscillators to "communicate"
through their phase relationships, leading to synchronized behavior and the emergence of meaning
through collective interactions.
The degree of synchronization:
R = |(1/N) Σ e^(iθ_j)|
Where R = 0 indicates complete desynchronization and R = 1 indicates perfect synchronization. The resonance index measures how effectively phase coupling has enabled oscillators to communicate and align, creating coherent collective behavior.
Phase coupling is the fundamental mechanism in the AION model that enables meaning to emerge:
- Oscillators "communicate" through their phase relationships
- Phase differences create coupling forces that pull oscillators toward alignment
- As oscillators synchronize, coherent patterns emerge from their interactions
- This mirrors how neurons communicate through synchronized firing to create perceptions
- The clustering and synchronization you observe represents the emergence of meaning through phase-based communication
While phase coupling shares some similarities with reinforcement learning, there are important differences:
Similarities:
- Both use feedback mechanisms to adjust behavior over time
- Both converge to stable or optimal states
- Both produce emergent behavior from interactions
Key Differences:
| Aspect | Phase Coupling | Reinforcement Learning |
|---|---|---|
| Rewards | Implicit (energy minimization) | Explicit (reward signals) |
| Nature | Deterministic (differential equations) | Often stochastic (exploration) |
| Scope | Collective (all oscillators interact) | Typically individual (agent-environment) |
| Process | Continuous phase updates | Discrete actions and states |
| Exploration | No exploration needed | Requires exploration |
| Type | Physics-based self-organization | Algorithm-based learning |
In the AION Context: Phase coupling is more similar to:
- Gradient descent: Moving toward lower energy states
- Self-organization: Emergent order from local interactions
- Collective intelligence: Group behavior emerging from individual interactions
Unlike reinforcement learning, phase coupling doesn't require explicit rewards or exploration—the system naturally evolves toward synchronized states through the physics of phase interactions, representing a form of emergent meaning rather than goal-directed learning.
While the AION model shares some mathematical similarities with diffusion models (used in generative AI), there are important differences:
Similarities:
- Both use differential equations to model dynamics over time
- Both can show transitions from disorder to order
- Both can involve stochastic processes (though AION is often deterministic)
- Both can be viewed as optimization processes
Key Differences:
| Aspect | AION/Kuramoto | Diffusion Models |
|---|---|---|
| Purpose | Synchronization | Generative (create new data) |
| Process | Disorder → Order (synchronization) | Data → Noise → Denoise → New Data |
| Goal | Achieve phase alignment | Learn data distribution |
| Output | Synchronized oscillators | Generated images/text/etc. |
| Nature | Often deterministic | Stochastic (noise-driven) |
| Learning | No learning (physics-based) | Requires training on data |
| Mechanism | Phase coupling | Noise diffusion and denoising |
In the AION Context: The AION model is more similar to:
- Reaction-diffusion systems: Patterns emerge from local interactions
- Self-organization: Order emerges from collective dynamics
- Gradient flows: Moving toward stable states (energy minimization)
Unlike diffusion models, AION is not generative—it doesn't create new data. Instead, it focuses on synchronization and the emergence of meaning through phase alignment, representing a form of physics-based self-organization rather than learned data generation.
The AION model's phase coupling mechanism has potential applications in AI systems, particularly in Retrieval-Augmented Generation (RAG) workflows:
1. Document Embedding Synchronization
- Represent documents as oscillators with phases derived from embeddings
- Phase coupling aligns semantically similar documents through synchronization
- Clustered documents form coherent retrieval groups, improving search quality
2. Query-Document Alignment
- Use phase coupling to align query representations with relevant document representations
- Synchronization strength based on semantic similarity improves relevance matching
- Better alignment leads to more accurate retrieval
3. Multi-Modal Binding
- Bind text, images, and other modalities through phase synchronization
- Similar to neuro-symbolic binding, different modalities synchronize to form unified representations
- Enables coherent multi-modal RAG systems
4. Context Aggregation
- Synchronize multiple retrieved contexts before generation
- Phase coupling creates coherent aggregated context from multiple sources
- Reduces contradictions and improves generation quality
- Representation: Map embeddings/vectors to oscillator phases (e.g., use embedding angles or projections)
- Frequency Assignment: Assign intrinsic frequencies to documents representing their baseline importance or relevance
- Coupling: Define coupling strength based on similarity, relevance, or learned relationships
- Synchronization: Let phases align through Kuramoto dynamics
- Emergent Meaning: Use synchronized clusters as coherent representations for retrieval or generation
When nodes represent documents in a RAG system, frequency (ω) takes on a specific meaning:
What Frequency Represents:
- Intrinsic Relevance/Importance: A document's baseline importance or relevance score that exists independently of queries or other documents
- Query-Independent Activity: How "active" or prominent a document is in the knowledge base
- Domain-Specific Centrality: How central a document is to its domain
Possible Interpretations:
- Citation-based: Documents with more citations have higher frequency
- Recency-based: More recent documents have higher frequency
- Authority-based: Documents from authoritative sources have higher frequency
- Learned Importance: Frequency learned from usage patterns or relevance feedback
How It Works:
- Each document gets an intrinsic frequency (ω) representing its baseline importance
- Documents with higher frequency have more "influence" in the synchronization process
- Phase coupling still aligns documents based on semantic similarity
- Synchronization clusters semantically similar documents, with important documents having more weight
- Documents with different importance levels can still synchronize if semantically similar
Example:
- Document A (high frequency, ω=2.0): Important research paper with many citations
- Document B (low frequency, ω=0.5): Supporting material or less authoritative source
- With coupling: They can still synchronize if semantically similar, but the important document has more influence in the cluster
In RAG systems, frequency represents each document's intrinsic importance or relevance, independent of queries or other documents, while phase coupling handles the interactions and synchronization between documents.
- Attention Mechanisms: Model attention as synchronization between query and key representations
- Semantic Clustering: Use phase synchronization to cluster semantically related concepts
- Knowledge Graph Alignment: Align entities and relations through phase coupling
- Multi-Agent Coordination: Coordinate multiple AI agents through phase synchronization
- Embedding-to-Phase Mapping: Convert high-dimensional embeddings to phases (e.g., using angles, projections, or learned mappings)
- Coupling Definition: Define meaningful coupling between representations (similarity-based, learned, or task-specific)
- Scale: Handle large numbers of documents/embeddings efficiently (may require approximations or hierarchical approaches)
- Integration: Combine with existing RAG pipelines (embedding models, vector databases, LLMs)
In RAG workflows, phase coupling could:
- Improve Retrieval: Clustering related documents through synchronization
- Enhance Context Understanding: Creating synchronized representations from multiple sources
- Enable Multi-Modal Binding: Binding different data types through phase alignment
- Create Coherent Aggregation: Forming unified context from diverse sources before generation
The phase coupling mechanism represents a novel approach to creating coherent, synchronized representations from diverse information sources—a key challenge in modern AI systems.
[Add license information here]
[Add contributing guidelines here]