# Knowledge Graphs ## Why Knowledge Graphs - Knowledge representation brings the ability to represent entities and relations with high reliability, explainability, and reusability. - Opportunities in machine learning: - Data augmentation - Zero-shot learning by deduction from a knowledge graph - Interpretability ## Applications - Question Answering - Storing information - Recommendation System (Netflix uses in production) - Supply chain management ## Challenges - No universal definition of KG and set of best practices - Knowledge integration - Capturing evolving nature of knowledge - Evaluation --- ## References 1. Knowledge Graph: The Perfect Complement to Machine Learning https://towardsdatascience.com/knowledge-graph-bb78055a7884