Guest Editorial Special Issue on Intrinsically Motivated Open-Ended Learning (IMOL)

The concept of lifelong
[1]
, continual
[2]
, progressive
[3]
, or open-ended
[4]
,
[5]
,
[6]
learning by artificial agents or robots is of interest to researchers because it permits robots to adapt to multiple tasks over the course of their life and progressively accumulate knowledge
[1]
,
[3]
,
[7]
. This means that the agent or robot is less likely to become obsolete due to environmental changes, and it may be better equipped to respond to new tasks through the accumulated knowledge. This can include motor skills
[8]
, high-level behaviors
[9]
, vision
[10]
, and social skills
[11]
, among others. On the other hand, the ability to model and thus “understand” lifelong learning is of potential relevance for robots that interact with children or the elderly, or software agents that interact with people in education or training settings.