This paper introduces the concept of Probabilistic Relational Models (PRMs) and discusses methods for learning them from relational databases. PRMs extend traditional Bayesian networks by allowing properties of an object to depend probabilistically on other properties of that object and on properties of related objects. The authors present techniques for parameter estimation and structure learning, including the use of standard database retrieval techniques to efficiently handle large datasets. They also describe a heuristic search algorithm for finding high-scoring structures and provide experimental results on both real and synthetic relational databases. The paper highlights the challenges and potential extensions of PRMs, such as handling missing data, hidden variables, and automated discovery of hidden variables.This paper introduces the concept of Probabilistic Relational Models (PRMs) and discusses methods for learning them from relational databases. PRMs extend traditional Bayesian networks by allowing properties of an object to depend probabilistically on other properties of that object and on properties of related objects. The authors present techniques for parameter estimation and structure learning, including the use of standard database retrieval techniques to efficiently handle large datasets. They also describe a heuristic search algorithm for finding high-scoring structures and provide experimental results on both real and synthetic relational databases. The paper highlights the challenges and potential extensions of PRMs, such as handling missing data, hidden variables, and automated discovery of hidden variables.