Advertisement

Reducing Training Environments in Evolutionary Robotics Through Ecological Modularity

  • Collin CappelleEmail author
  • Anton Bernatskiy
  • Josh Bongard
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 10384)

Abstract

Due to the large number of evaluations required, evolutionary robotics experiments are generally conducted in simulated environments. One way to increase the generality of a robot’s behavior is to evolve it in multiple environments. These environment spaces can be defined by the number of free parameters (f) and the number of variations each free parameter can take (n). Each environment space then has \(n^f\) individual environments. For a robot to be fit in the environment space it must perform well in each of the \(n^f\) environments. Thus the number of environments grows exponentially as n and f are increased. To mitigate the problem of having to evolve a robot in each environment in the space we introduce the concept of ecological modularity. Ecological modularity is here defined as the robot’s modularity with respect to free parameters in its environment space. We show that if a robot is modular along m of the free parameters in its environment space, it only needs to be evolved in \(n^{f-m+1}\) environments to be fit in all of the \(n^f\) environments. This work thus presents a heretofore unknown relationship between the modularity of an agent and its ability to generalize evolved behaviors in new environments.

Notes

Acknowledgments

We like to acknowledge financial support from NSF awards PECASE-0953837 and INSPIRE-1344227 as well as the Army Research Office contract W911NF-16-1-0304. We also acknowledge computation provided by the Vermont Advanced Computing Core.

References

  1. 1.
    Bongard, J., Bernatskiy, A., Livingston, K., Livingston, N., Long, J., Smith, M.: Evolving robot morphology facilitates the evolution of neural modularity and evolvability. In: Proceedings of the 2015 Genetic and Evolutionary Computation Conference, p. 129136. ACM, Madrid (2015)Google Scholar
  2. 2.
    Bongard, J.: Morphological change in machines accelerates the evolution of robust behavior. Proc. Nat. Acad. Sci. 108(4), 1234–1239 (2011)CrossRefGoogle Scholar
  3. 3.
    Bongard, J.C.: Spontaneous evolution of structural modularity in robot neural network controllers. In: Proceedings of the 2011 Genetic and Evolutionary Computation Conference, pp. 251–258. ACM, Dublin (2011)Google Scholar
  4. 4.
    Cappelle, C.K., Bernatskiy, A., Livingston, K., Livingston, N., Bongard, J.: Morphological modularity can enable the evolution of robot behavior to scale linearly with the number of environmental features. Front. Rob. AI 3, 59 (2016)Google Scholar
  5. 5.
    Clune, J., Mouret, J.B., Lipson, H.: The evolutionary origins of modularity. Proc. R. Soc. B Biol. Sci. 280(1755), 20122863 (2013)CrossRefGoogle Scholar
  6. 6.
    Ellefsen, K.O., Mouret, J.B., Clune, J.: Neural modularity helps organisms evolve to learn new skills without forgetting old skills. PLoS Comput. Biol. 11(4), e1004128 (2015)CrossRefGoogle Scholar
  7. 7.
    Espinosa-Soto, C., Wagner, A.: Specialization can drive the evolution of modularity. PLoS Comput. Biol. 6(3), e1000719 (2010)MathSciNetCrossRefGoogle Scholar
  8. 8.
    French, R.M.: Catastrophic forgetting in connectionist networks. Trends Cogn. Sci. 3(4), 128–135 (1999)CrossRefGoogle Scholar
  9. 9.
    Gruau, F.: Automatic definition of modular neural networks. Adapt. Behav. 3, 151–183 (1994)CrossRefGoogle Scholar
  10. 10.
    Jakobi, N., Husbands, P., Harvey, I.: Noise and the reality gap: the use of simulation in evolutionary robotics. In: Morán, F., Moreno, A., Merelo, J.J., Chacón, P. (eds.) ECAL 1995. LNCS, vol. 929, pp. 704–720. Springer, Heidelberg (1995). doi: 10.1007/3-540-59496-5_337 CrossRefGoogle Scholar
  11. 11.
    Kashtan, N., Alon, U.: Spontaneous evolution of modularity and network motifs. PNAS 102(39), 13773 (2005)CrossRefGoogle Scholar
  12. 12.
    Lehman, J., Risi, S., D’Ambrosio, D., Stanley, K.O.: Encouraging reactivity to create robust machines. Adapt. Behav. 21, 484–500 (2013)CrossRefGoogle Scholar
  13. 13.
    Lipson, H., Pollack, J.B., Suh, N.P., Wainwright, P.: On the origin of modular variation. Evolution 56(8), 1549–1556 (2002)CrossRefGoogle Scholar
  14. 14.
    Matarić, M., Cliff, D.: Challenges in evolving controllers for physical robots. Rob. Auton. Syst. 19(1), 67–83 (1996)CrossRefGoogle Scholar
  15. 15.
    Newman, M.E., Girvan, M.: Finding and evaluating community structure in networks. Phys. Rev. E 69(2), 026113 (2004)CrossRefGoogle Scholar
  16. 16.
    Schmidt, M., Lipson, H.: Age-fitness pareto optimization. In: Riolo, R., McConaghy, T., Vladislavleva, E. (eds.) Genetic Programming Theory and Practice VIII. Genetic and Evolutionary Computation, vol. 8, pp. 129–146. Springer, New York (2011). doi: 10.1007/978-1-4419-7747-2_8 CrossRefGoogle Scholar
  17. 17.
    Wagner, G., Pavlicev, M., Cheverud, J.: The road to modularity. Nat. Rev. Genetics 8(12), 921–931 (2007)CrossRefGoogle Scholar
  18. 18.
    Yamashita, Y., Tani, J.: Emergence of functional hierarchy in a multiple timescale neural network model: a humanoid robot experiment. PLoS Comput. Biol. 4(11), e1000220 (2008)CrossRefGoogle Scholar

Copyright information

© Springer International Publishing AG 2017

Authors and Affiliations

  • Collin Cappelle
    • 1
    Email author
  • Anton Bernatskiy
    • 1
  • Josh Bongard
    • 1
  1. 1.The University of VermontBurlingtonUSA

Personalised recommendations