Information Particle System

Version: 2.0.0
Author: Beijing Qiuyishusheng Technology Center
License: MIT

Overview

Information Particle System is a framework for information structure analysis and quality assessment. The system discretizes continuous information into interpretable particles, each characterized by 12-dimensional transparent features.

Core Concepts

Information Particle

The fundamental unit of information representation, analogous to pixels in image processing. Each particle contains:

• 12-dimensional feature vector
• Structure-Information-Function (SIF) quality metric
• Original content for lossless reconstruction

Feature Structure

[Time Dimension - 4D]
├── inner_time: Relative temporal position
├── outer_time: Absolute timestamp
├── time_flow: Rate of change
└── current_time: Current state marker

[Spatial Dimension - 3D]
├── spatial_x: Position coordinate
├── spatial_y: Content mean value
└── spatial_z: Content variance

[Structure Dimension - 4D]
├── density: Information density
├── connectivity: Inter-particle correlation
├── stability: Content stability
└── energy: Information magnitude

[Quality Metric - 1D]
└── sif_value: Composite quality score

SIF Metric

Structure-Information-Function (SIF) value quantifies information quality:

SIF = 0.3 × Structure + 0.5 × Information + 0.2 × Function

Range: [0, 1], where higher values indicate richer information structure.

Key Features

• Lossless Reconstruction: Perfect reconstruction with MSE=0
• Transparent Features: 12 interpretable dimensions
• Quality Assessment: Automated SIF scoring
• Topological Mapping: Spherical coordinate representation
• Pure Mathematical: Rule-based, no neural networks

Installation

git clone https://github.com/changsheng137/information-sphere-system.git
cd information-sphere-system
pip install -r requirements.txt

Quick Start

from src.information_oriented_system import InformationOrientedSystemV2
import torch

# Initialize system
system = InformationOrientedSystemV2(particle_size=4)

# Process data
data = torch.randn(28, 28)
output = system.forward(data)

# Results
print(f"Particles: {output['num_particles']}")
print(f"Groups: {output['num_groups']}")
print(f"SIF Score: {output['avg_sif']:.4f}")

# Lossless reconstruction
reconstructed = system.reconstruct(output)
mse = torch.nn.functional.mse_loss(reconstructed, data)
print(f"MSE: {mse:.10f}")  # Expected: 0.0

Architecture

Input Data [H×W]
    ↓
┌─────────────────────────┐
│ Particle Extraction     │
│ ├── Partitioning        │
│ ├── Feature Calculation │
│ └── SIF Computation     │
└─────────────────────────┘
    ↓
┌─────────────────────────┐
│ Group Construction      │
│ ├── Similarity Analysis │
│ └── Clustering          │
└─────────────────────────┘
    ↓
┌─────────────────────────┐
│ Spherical Mapping       │
│ └── Topology Projection │
└─────────────────────────┘
    ↓
┌─────────────────────────┐
│ Lossless Reconstruction │
└─────────────────────────┘

Experimental Validation

Reconstruction Quality

• Test samples: 350+ images
• Perfect reconstruction: 100% (MSE=0)
• Datasets: MNIST, Fashion-MNIST

Baseline Comparison

Method	MSE	Processing Time	Additional Info
Direct Storage	0.0	0.03ms	None
PCA (64)	0.0051	0.34ms	None
PCA (128)	0.0009	0.10ms	None
Ours	0.0	1.00ms	12D + SIF + Structure

Component Validation

Ablation study confirms contribution of each component:

• Full system: Complete feature set
• Without SIF: No quality assessment
• Simplified features: Incomplete description
• Without grouping: No semantic structure
• Without connectivity: No topological information

Applications

1. Quality Assessment

quality = system.forward(image)['avg_sif']
# Interpret: 0.6+ high, 0.4-0.6 medium, <0.4 low

2. Anomaly Detection

anomalies = [p for p in particles if p.sif_value < threshold]

3. Structure Visualization

# Generate SIF distribution, connectivity network, sphere projection

Performance

Metric	Value
Processing Speed	~1ms per 28×28 image
Memory Overhead	+10.8% (12D features)
Time Complexity	O(n + k²)
Space Complexity	O(k × (12 + s))

Documentation

• Theoretical Foundation
• Experiment Report
• User Guide
• API Reference

Project Structure

information-sphere-v1.0/
├── src/
│   ├── information_particle_system.py    # Core system
│   └── information_oriented_system.py    # Main interface
├── experiments/
│   ├── baseline_comparison.py            # Baseline tests
│   └── ablation_study.py                 # Component validation
├── examples/
│   ├── basic_usage.py                    # Quick start
│   └── application_demo.py               # Application examples
├── docs/                                 # Documentation
└── tests/                                # Unit tests

Citation

If you use this system in your research, please cite:

@software{qiuyishusheng2025information,
  title={Information Particle System: A Framework for Information Structure Analysis},
  author={Beijing Qiuyishusheng Technology Center},
  year={2025},
  version={2.0.0}
}

License

MIT License - see LICENSE file

Contact

• GitHub: @changsheng137
• Issues: GitHub Issues

Acknowledgments

Theoretical foundations:

• Shannon's Information Theory
• Cognitive Dimension Theory
• Time-Set Dimension Framework

---

Beijing Qiuyishusheng Technology Center © 2025