56 lines
1.5 KiB
Markdown
56 lines
1.5 KiB
Markdown
# 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
|