# Psycho-Symbolic Reasoner - Cognitive Bias Modeling

## Overview
Hybrid reasoning combining psychological models (cognitive biases, heuristics) with symbolic logic and subsymbolic patterns.

## Purpose
Model human-like reasoning including systematic biases, demonstrating more realistic AI decision-making.

## Operations
- **Psychological Models**: 3 (confirmation bias, availability heuristic, anchoring)
- **Symbolic Rules**: 2 logical inference rules
- **Subsymbolic Patterns**: 5 neural activation patterns
- **Hybrid Reasoning**: 5 integrated decisions

## Results
- **Throughput**: 2.04 ops/sec
- **Latency**: 479ms avg
- **Memory**: 23 MB
- **Psychological Models**: 3
- **Symbolic Rules**: 2
- **Subsymbolic Patterns**: 5
- **Hybrid Instances**: 5

## Technical Details

### Psychological Layer
```typescript
model: 'confirmation_bias'
strength: 0.88
// Tendency to favor confirming evidence
```

### Symbolic Layer
```typescript
rule: 'IF bias_detected THEN adjust_confidence'
confidence: 0.92
```

### Integration
Detects cognitive biases → Applies corrective symbolic rules → Uses subsymbolic patterns for nuanced decisions

## Applications
- **Decision Support Systems**: Bias-aware recommendations
- **Educational Tools**: Teaching critical thinking
- **UX Design**: Predict user behavior patterns
- **Negotiation AI**: Model human decision-making

## Cognitive Biases Modeled
1. Confirmation bias
2. Availability heuristic
3. Anchoring effect
4. Representativeness heuristic
5. Framing effects

**Status**: ✅ Operational | **Package**: psycho-symbolic-reasoner