Evaluation of Internal Models in Autonomous Learning

S. C. Smith, J. M. Herrmann

Research output: Contribution to journalArticlepeer-review


Internal models (IMs) can represent relations between sensors and actuators in natural and artificial agents. In autonomous robots, the adaptation of IMs and the adaptation of the behaviour are interdependent processes which have been studied under paradigms for self-organisation of behaviour such as homeokinesis. We compare the effect of various types of IMs on the generation of behaviour in order to evaluate model quality across different behaviours. The considered IMs differ in the degree of flexibility and expressivity related to, respectively, learning speed and structural complexity of the model. We show that the different IMs generate different error characteristics which in turn lead to variations of the self-generated behaviour of the robot. Due to the trade-off between error minimisation and complexity of the explored environment, we compare the models in the sense of Pareto optimality. Among the linear and nonlinear models that we analyse, echo-state networks achieve a particularly high performance which we explain as a result of the combination of fast learning and complex internal dynamics. More generally, we provide evidence that Pareto optimisation is preferable in autonomous learning as it allows that a special solution can be negotiated in any particular environment.
Original languageEnglish
Pages (from-to)463 - 472
Number of pages10
JournalIEEE Transactions on Cognitive and Developmental Systems
Issue number4
Early online date22 Aug 2018
Publication statusPublished - 1 Dec 2019


  • Robot sensing systems
  • Adaptation models
  • Predictive models
  • Complexity theory
  • Mathematical model
  • Biological system modeling
  • Autonomous robot
  • internal model
  • prediction error
  • homeokinesis
  • time-loop error.


Dive into the research topics of 'Evaluation of Internal Models in Autonomous Learning'. Together they form a unique fingerprint.

Cite this