2c1acda2f3
- Update DQN trainer with gradient collapse detection warmup - Add portfolio tracker improvements - Include hyperopt trial results (multiple Sharpe ratio experiments) - Add new test files for action/position sign convention, early stopping, cash reserve bugs, and portfolio execution - Update trained model files - Add Claude Code configuration and skills 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
1.5 KiB
1.5 KiB
Neural Pattern Training
Purpose
Continuously improve coordination through neural network learning.
How Training Works
1. Automatic Learning
Every successful operation trains the neural networks:
- Edit patterns for different file types
- Search strategies that find results faster
- Task decomposition approaches
- Agent coordination patterns
2. Manual Training
Tool: mcp__claude-flow__neural_train
Parameters: {
"pattern_type": "coordination",
"training_data": "successful task patterns",
"epochs": 50
}
3. Pattern Types
Cognitive Patterns:
- Convergent: Focused problem-solving
- Divergent: Creative exploration
- Lateral: Alternative approaches
- Systems: Holistic thinking
- Critical: Analytical evaluation
- Abstract: High-level design
4. Improvement Tracking
Tool: mcp__claude-flow__neural_status
Result: {
"patterns": {
"convergent": 0.92,
"divergent": 0.87,
"lateral": 0.85
},
"improvement": "5.3% since last session",
"confidence": 0.89
}
Pattern Analysis
Tool: mcp__claude-flow__neural_patterns
Parameters: {
"action": "analyze",
"operation": "recent_edits"
}
Benefits
- 🧠 Learns your coding style
- 📈 Improves with each use
- 🎯 Better task predictions
- ⚡ Faster coordination
CLI Usage
# Train neural patterns via CLI
npx claude-flow neural train --type coordination --epochs 50
# Check neural status
npx claude-flow neural status
# Analyze patterns
npx claude-flow neural patterns --analyze