InferIQ is a powerful application that combines deductive reasoning, Bayesian networks, probabilistic programming, and machine learning to provide accurate inferences with confidence scores.
- Data & Rule Management: Upload and manage data files and define reasoning rules
- Inference & Validation: Run inference on your data using multiple reasoning methods
- Hyper-Parameter Tuning: Optimize the weights of different reasoning methods
- User Management: Manage users with different roles and permissions
- NLI: Supports natural language to SWRL conversion
The system now supports natural language rule creation with:
- Natural language to SWRL conversion using fine-tuned LLMs
- Ontology-aware validation
- Interactive feedback for rule correction
- Bidirectional SWRL/Natural Language views
Edit the .env file to configure the application:
# Storage path for all file storage
INFERIQ_STORAGE=/path/to/your/storage
- Login System: Detailed documentation of the login system.
- Load your ontology in the Ontology Management tab
- Switch to "Natural Language" input mode in Rule Management
- Describe your rule in plain English
- Click "Convert to SWRL" to generate formal rules
- Review and edit generated SWRL if needed
- Validate against your ontology before applying
- User Input:
Senior customers get 10% discounts
- LLM Processing:
Customer(?c) ∧ hasAge(?c, ?a) ∧ greaterThanOrEqual(?a, 65) → applyDiscount(?c, 0.1)
- Validation:
Error: applyDiscount not in ontology
- Feedback:
Found issues:
- Undefined property: applyDiscount. Did you mean: hasDiscount, discountRate?
- User Correction:
Use hasDiscount instead
- Final Output:
Customer(?c) ∧ hasAge(?c, ?a) ∧ swrlb:greaterThanOrEqual(?a, 65) → hasDiscount(?c, 0.1)