by-causal-reasoning

What problem does it solve?

Translates discriminating diagnostic features into canonical, evidence-grounded mechanistic hypotheses anchored in a knowledge graph, enabling auditable explanations for design failures.

Core Features & Use Cases

• Maps discriminating features to canonical mechanisms using a knowledge graph, providing traceable rationale.
• Generates up to five hypotheses, each with structured supporting and contradicting evidence, and a falsifiable prediction.
• Provides downstream guidance for campaigns (e.g., optimization, hypothesis debate, or knowledge-store updates) based on confidence and evidence mix.

Quick Start

Run the generate_hypotheses.py workflow with a diagnosis.json and knowledge graph data to produce hypotheses.json for downstream analysis.