Master-apprentice approach

by Mika Hämäläinen

Also available as: PDF, LaTeX.


Master-apprentice approach is a method in computational creativity research which consists of at least two computational agents: master and apprentice [1]. In the literature, genetic algorithms have been used as a master and recurrent neural networks as an apprentice [2].
The idea of the approach is that a computationally creative and interpretable master can produce training data to a black box neural network model. This training data can be enhanced with human authored data as well [3].
  • [1]^ Alnajjar, K., & Hämäläinen, M. (2018). A master-apprentice approach to automatic creation of culturally satirical movie titles. In Proceedings of the 11th International Conference on Natural Language Generation (pp. 274-283).
  • [2]^ Hämäläinen, M., & Alnajjar, K. (2019). Let’s FACE it. Finnish Poetry Generation with Aesthetics and Framing. In Proceedings of the 12th International Conference on Natural Language Generation (pp. 290-300).
  • [3]^ Hämäläinen, M., & Alnajjar, K. (2019). Modelling the Socialization of Creative Agents in a Master-Apprentice Setting: The Case of Movie Title Puns. In The 10th International Conference on Computational Creativity (pp. 266-273). Association for Computational Creativity.
Alternative definitions:
Ontology information:
Master-apprentice approach
Related papers: