# MAYA Self-Evolution v2

Autonomous self-improvement system for steganography detection using genetic algorithms.

## Overview

MAYA Self-Evolution v2 is an autonomous system that continuously improves steganography detection capabilities through genetic algorithm-based evolution. The system optimizes detection weights, thresholds, and feature importance to maximize detection accuracy.

## Features

- **Genetic Algorithm Evolution**: Population-based optimization with selection, crossover, and mutation
- **Fitness-Based Selection**: Elite preservation ensures best performing genomes propagate
- **Diversity Preservation**: Maintains population diversity to avoid premature convergence
- **Convergence Detection**: Automatically stops when improvement stagnates
- **Multi-Feature Optimization**: Simultaneously optimizes 8 steganalysis features

## Supported Features

1. `lsb_variance` - LSB variance analysis
2. `alpha_entropy` - Alpha channel entropy
3. `dct_coefficients` - DCT coefficient analysis
4. `histogram_analysis` - Histogram analysis
5. `spatial_correlation` - Spatial correlation
6. `bit_plane_analysis` - Bit plane analysis
7. `chi_square` - Chi-square test
8. `rs_analysis` - RS analysis

## Usage

### Python API

```python
from maya_evolution import MAYASelfEvolution, create_maya_evolution_system

# Initialize system
mayan = create_maya_evolution_system()

# Define test results (detection confidence per feature)
test_results = {
    "lsb_variance": 0.85,
    "alpha_entropy": 0.72,
    "dct_coefficients": 0.68,
    "histogram_analysis": 0.91,
    "spatial_correlation": 0.45,
    "bit_plane_analysis": 0.77,
    "chi_square": 0.83,
    "rs_analysis": 0.69
}

# Run evolution
for gen in range(10):
    state = mayan.evolve_generation(test_results)
    print(f"Gen {state.generation}: Best={state.best_fitness:.4f}")

# Get optimized genome
best = mayan.get_best_genome()
print(f"Best fitness: {best.fitness}")
print(f"Optimized weights: {best.detection_weights}")
```

### Web Interface

Open `maya_evolution.html` in a browser to see the interactive visualization:
- Real-time fitness tracking
- Feature importance charts
- Evolution controls

## Files

- `maya_evolution.py` - Core evolution engine
- `maya_evolution.html` - Web visualization
- `api.py` - Server-side API handler

## Technical Details

### Evolution Parameters

- **Population Size**: 20 genomes
- **Elite Size**: 4 best genomes preserved
- **Crossover Rate**: 70%
- **Mutation Rate**: 10%
- **Convergence Threshold**: 0.001

### Fitness Function

Fitness = Weighted detection score + Diversity bonus

Where diversity bonus promotes exploration of solution space.

## Security

All operations use only allowed imports:
- json, math, base64, hashlib, datetime, re, string
- itertools, collections, dataclasses, html, urllib.parse, random

No file system, network, or code execution operations.