Abstract

Once artificial agents are able to support human beings and their activities we will see a shift from these days cooperative activities to more elaborated teamwork settings. There, agents are not only assistants but indeed become team members. Yet, such integration causes significant challenges. One of these challenges is to make the behaviour of all team members predictable in order to account for an effective collaboration. Interestingly, this challenge is not well-defined and different involved research areas refer to it using different terms. This paper explains what is meant by predictability eventually proposing a definition for the challenge. In doing so, we discuss predictability referring to joint human-agent activities, the involved learning problem, and the inherent planning part, which was named human-aware planning recently.

@InProceedings{Ahrndt2016a,
  author       = {Sebastian Ahrndt AND Johannes F"ahndrich AND Sahin Albayrak},
  title        = {Human-Agent Teamwork: What is predictability, why is it important?},
  booktitle    = {Proceedings of the 31st Annual ACM Symposium on Applied Computing},
  year         = {2016},
  volume       = {1},
  series       = {SAC '16},
  organization = {ACM SIGAPP},
  publisher    = {ACM},
  month        = {April 4--8},
  pages        = {284--286},
  doi          = {10.1145/2851613.2851928},
  url          = {http://doi.acm.org/10.1145/2851613.2851928},
  address      = {New York, NY, USA},
  keywords     = {human-aware planning, joint activities, predictability, teamwork}
}