Evolving visually guided agents in an ambiguous virtual world

ABSTRACT

The fundamental challenge faced by any visual system within natural environments is the ambiguity caused by the fact that light that falls on the system's sensors conflates multiple attributes of the physical world. Understanding the computational principles by which natural systems overcome this challenge and generate useful behaviour remains the key objective in neuroscience and machine vision research. In this paper we introduce Mosaic World, an artificial life model that maintains the essential characteristics of natural visual ecologies, and which is populated by virtual agents that - through 'natural' selection - come to resolve stimulus ambiguity by adapting the functional structure of their visual networks according to the statistical structure of their ecological experience. Mosaic World therefore presents us with an important tool for exploring the computational principles by which vision can overcome stimulus ambiguity and usefully guide behaviour.

REFERENCES

• 1: Collins R. J. and Jefferson D. R. The evolution of sexual selection and female choice. In F. J. Varela and P. Bourgine, editors, Toward a Practice of Autonomous Systems: Proceedings of the First European Conference on Artificial Life, pages 327--336, Cambridge, MA, 1992.
• 2: Dyer, J. and Bentley, P. J. PLANTWORLD: Population Dynamics in Contrasting Environments. A late-breaking paper in Proceedings of the Genetic and Evolutionary Computation Conference (GECCO) 2002.
• 3: Dyer, J. R., Bentley, P. J. and Shah, P. PLANTWORLD: The Evolution of Plant Dormancy in Contrasting Environments. A late-breaking paper in Proceedings of the Genetic and Evolutionary Computation Conference (GECCO) 2001. pp. 67--74.
• 4: Rens Kortmann , Eric Postma , Jaap Van den Herik, Evolution of visual resolution constrained by a trade-off, Artificial Life, v.7 n.2, p.125-145, Spring 2001 [doi>10.1162/106454601753138970]
• 5: John R. Koza , Jonathan Roughgarden , James P. Rice, Evolution of food-foraging strategies for the Caribbean anolis lizard using genetic programming, Adaptive Behavior, v.1 n.2, p.171-199, Fall 1992 [doi>10.1177/105971239200100203]
• 6: Achim Liese , Daniel Polani , Thomas Uthmann, A study of the simulated evolution of the spectral sensitivity of visual agent receptors, Artificial Life, v.7 n.2, p.99-124, Spring 2001 [doi>10.1162/106454601753138961]
• 7: Lotto, R.B. and Purves, D. From the Cover: The effects of color on brightness. Nature Neuroscience 1999 2:1010--1014.
• 8: Lotto, R.B. and Purves, D. From the cover: An empirical explanation of color contrast. Proceedings of the National Academy of Science 2000 USA 97:12834--12839.
• 9: Lotto, R.B. and Purves, D. From the cover: An empirical explanation of the Chubb illusion. Journal of Cognitive Neuroscience 2001 13:547--555.
• 10: Purves, D.P. and Lotto, R.B. Why we see what we do: A wholly probabilistic strategy of vision. Sinaur Associates INC. (Sunderland Massachusetts) and Macmillan Press (London, UK), 2003.
• 11: D. Terzopoulos , T. F. Rabie, Animat vision: Active vision in artificial animals, Proceedings of the Fifth International Conference on Computer Vision, p.801, June 20-23, 1995
• 12: Unemi T., Kaneko Y. and Takahashi I. War and Peace among Artificial Nations - a model and simulation based on a two-layered multi-agent system, ECAL 2003, Springer-Verlag LNCS/LNAI series, LNAI 2801, pp. 146--153, 2003.

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
        I.2.10 Vision and Scene Understanding
            Subjects: Modeling and recovery of physical attributes

Additional Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
        I.2.10 Vision and Scene Understanding
            Subjects: Perceptual reasoning

General Terms:
Algorithms

Keywords:
ambiguous environments, mosaic world, vision evolution