Top Results for Commonsense AI
ConceptNet is a multilingual semantic network containing over 90 million everyday relationships extracted from text and structured to represent common sense knowledge about the world, enabling reasoning and inference tasks.
Why this score
Highly regarded open commonsense graph with multilingual reach and extensive research use; noisy relations, shallow context and coverage imbalance are established weaknesses.
Scoring methodology
Cyc is a structured knowledge base built through iterative refinement by a team of experts, aiming to represent common sense and everyday understanding in a machine-readable format with detailed relationships between concepts.
Why this score
Pioneering scale and ambition in symbolic common-sense knowledge; opacity, proprietary history, labor intensity and limited demonstrated general reasoning temper acclaim.
Scoring methodology
