The alternative to implement AI algorithm manual, is to use pre-programmed AI libraries. Most of them are released for the Python scripting language and are based on machine learning.[1] For Computer vision, also some libraries are available.[2]

An AI library is often bundled with lectures and books in which practical examples are introduced. The idea is to make normal AI courses easier to follow and to plot nice looking 3d charts to the screen. Libraries like Scikit-learn, matplotlib, Tensorflow and OpenCV are sometimes used in serious universities projects to test out new ideas with a low amount of programming effort.

The pybrain AI library has to be mentioned separate, because it was an early example (year 2010) for an AI library which integrates lots of concepts like reinforcement learning, LSTM networks and three layer networks into a single python API.[3] The more recent AI libraries, Tensorflow and Keras, were released later.


  1. EuroPython Conference: Ian Lewis - Deep Learning with Python & TensorFlow
  2. Top 20 Python libraries for Data Science
  3. Schaul, Tom, et al. "PyBrain." Journal of Machine Learning Research 11.Feb (2010): 743-746.
Community content is available under CC-BY-SA unless otherwise noted.