1.5 KiB
1.5 KiB
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
model: 'confirmation_bias'
strength: 0.88
// Tendency to favor confirming evidence
Symbolic Layer
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
- Confirmation bias
- Availability heuristic
- Anchoring effect
- Representativeness heuristic
- Framing effects
Status: ✅ Operational | Package: psycho-symbolic-reasoner