Commonsense Knowledge Bases are knowledge bases (or knowledge graphs) organized around the representation of data that everyone is expected to know (common sense). The types of facts that may be expected within a commonsense knowledge base could “lemons are sour” or “the sky is blue.” A fully-functioning representation and reasoning process to apply common sense is currently an unsolved AI problem. These types of facts and inferences are particularly powerful as they play a crucial role in distinguishing what incomplete information may be valuable for human purposes.
Practical applications of commonsense knowledge bases include the monitoring of social media for bullying remarks, data to power chat bots, and even programmatic monitoring of violations of nuclear test ban treaties.