Advanced search
2 files | 3.80 MB Add to list

You only look as much as you have to : using the free energy principle for active vision

Toon Van de Maele (UGent) , Tim Verbelen (UGent) , Ozan Catal (UGent) , Cedric De Boom (UGent) and Bart Dhoedt (UGent)
Author
Organization
Project
Abstract
Active vision considers the problem of choosing the optimal next viewpoint from which an autonomous agent can observe its environment. In this paper, we propose to use the active inference paradigm as a natural solution to this problem, and evaluate this on a realistic scenario with a robot manipulator. We tackle this problem using a generative model that was learned unsupervised purely from pixel-based observations. We show that our agent exhibits information-seeking behavior, choosing viewpoints of regions it has not yet observed. We also show that goal-seeking behavior emerges when the agent has to reach a target goal, and it does so more efficiently than a systematic grid search.

Downloads

  • (...).pdf
    • full text (Accepted manuscript)
    • |
    • UGent only
    • |
    • PDF
    • |
    • 3.20 MB
  • 7781 i.pdf
    • full text (Accepted manuscript)
    • |
    • open access
    • |
    • PDF
    • |
    • 595.65 KB

Citation

Please use this url to cite or link to this publication:

MLA
Van de Maele, Toon, et al. “You Only Look as Much as You Have to : Using the Free Energy Principle for Active Vision.” Active Inference, First International Workshop, IWAI 2020, Co-Located with ECML/PKDD 2020, edited by Tim Verbelen et al., vol. 1326, Springer, 2020, pp. 92–100, doi:10.1007/978-3-030-64919-7_11.
APA
Van de Maele, T., Verbelen, T., Catal, O., De Boom, C., & Dhoedt, B. (2020). You only look as much as you have to : using the free energy principle for active vision. In T. Verbelen, P. Lanillos, C. Buckley, & C. De Boom (Eds.), Active inference, First International Workshop, IWAI 2020, Co-located with ECML/PKDD 2020 (Vol. 1326, pp. 92–100). online: Springer. https://doi.org/10.1007/978-3-030-64919-7_11
Chicago author-date
Van de Maele, Toon, Tim Verbelen, Ozan Catal, Cedric De Boom, and Bart Dhoedt. 2020. “You Only Look as Much as You Have to : Using the Free Energy Principle for Active Vision.” In Active Inference, First International Workshop, IWAI 2020, Co-Located with ECML/PKDD 2020, edited by Tim Verbelen, P. Lanillos, C. Buckley, and Cedric De Boom, 1326:92–100. Springer. https://doi.org/10.1007/978-3-030-64919-7_11.
Chicago author-date (all authors)
Van de Maele, Toon, Tim Verbelen, Ozan Catal, Cedric De Boom, and Bart Dhoedt. 2020. “You Only Look as Much as You Have to : Using the Free Energy Principle for Active Vision.” In Active Inference, First International Workshop, IWAI 2020, Co-Located with ECML/PKDD 2020, ed by. Tim Verbelen, P. Lanillos, C. Buckley, and Cedric De Boom, 1326:92–100. Springer. doi:10.1007/978-3-030-64919-7_11.
Vancouver
1.
Van de Maele T, Verbelen T, Catal O, De Boom C, Dhoedt B. You only look as much as you have to : using the free energy principle for active vision. In: Verbelen T, Lanillos P, Buckley C, De Boom C, editors. Active inference, First International Workshop, IWAI 2020, Co-located with ECML/PKDD 2020. Springer; 2020. p. 92–100.
IEEE
[1]
T. Van de Maele, T. Verbelen, O. Catal, C. De Boom, and B. Dhoedt, “You only look as much as you have to : using the free energy principle for active vision,” in Active inference, First International Workshop, IWAI 2020, Co-located with ECML/PKDD 2020, online, 2020, vol. 1326, pp. 92–100.
@inproceedings{8675420,
  abstract     = {{Active vision considers the problem of choosing the optimal next viewpoint from which an autonomous agent can observe its environment. In this paper, we propose to use the active inference paradigm as a natural solution to this problem, and evaluate this on a realistic scenario with a robot manipulator. We tackle this problem using a generative model that was learned unsupervised purely from pixel-based observations. We show that our agent exhibits information-seeking behavior, choosing viewpoints of regions it has not yet observed. We also show that goal-seeking behavior emerges when the agent has to reach a target goal, and it does so more efficiently than a systematic grid search.}},
  author       = {{Van de Maele, Toon and Verbelen, Tim and Catal, Ozan and De Boom, Cedric and Dhoedt, Bart}},
  booktitle    = {{Active inference, First International Workshop, IWAI 2020, Co-located with ECML/PKDD 2020}},
  editor       = {{Verbelen, Tim and Lanillos, P. and Buckley, C. and De Boom, Cedric}},
  isbn         = {{9783030649197}},
  issn         = {{1865-0929}},
  language     = {{eng}},
  location     = {{online}},
  pages        = {{92--100}},
  publisher    = {{Springer}},
  title        = {{You only look as much as you have to : using the free energy principle for active vision}},
  url          = {{http://dx.doi.org/10.1007/978-3-030-64919-7_11}},
  volume       = {{1326}},
  year         = {{2020}},
}

Altmetric
View in Altmetric