Estimation of distribution algorithms (EDAs) are a class of evolutionary algorithms which can be regarded as abstraction of genetic algorithms (GAs) because in the design of EDAs, the population, one of the GA distinctive features, is replaced by probabilistic models/distributions. Building and sampling from the models substitute for the common genetic operators, such as crossover and mutation. Due to their excellent optimization performance, EDAs have been intensively studied and extensively applied in recent years. In order to interest more people to join the research of EDAs, this paper plays as an entry level introduction to EDAs. It starts with introducing the origination and basic ideas of EDAs, followed by presenting the current EDA frameworks, which are broadly applied in many scientific and engineering disciplines. Finally, this paper also describes some ongoing topics and potential directions in the hope that readers may get further insights into EDAs.