Sensors and effectors determine how events in the world at large are related to the internal informational states of organisms and robotic devices. Sensors determine what kinds of distinctions (perceptual categories, features, primitives) can be made on the environment. By "evolving the sensors" perceptual repertoires can be adaptively altered and/or enlarged. To the extent that devices can adaptively choose their own feature primitives for themselves, they gain a greater measure of "epistemic autonomy" vis-a-vis their designers. Such devices are useful in ill-defined situations where the designer does not know a priori what feature primitives are adequate or optimum for solving a particular task. Several general strategies for adaptively altering or augmenting sensor function are proposed: 1) prosthesis: adaptive fabrication of new front-ends for existing sensors (e.g. telescopes), 2) active sensing: using motor-actions to alter what is sensed through interaction (poking, pushing, bending), 3) sensory evolution: adaptive construction of entirely new sensors (adaptive antibody construction, Gordon Pask's electrochemical device) and 4) internalized sensing: "bringing the world into the device" by creating internal, analog representations of the world out of which internal sensors extract newly-relevant properties (perceptual learning). Since many neural sensory representations appear to be analog and iconic in nature, neural assemblies can be adaptively formed to function as internal sensors that can switch behavior according to new perceptual categories.
[1]
John Churcher.
Implications and Applications of Piaget’s Sensorimotor Concepts
,
1984
.
[2]
William R. Uttal,et al.
An autocorrelation theory of form detection
,
1975
.
[3]
Robert L. Goldstone,et al.
The development of features in object concepts
,
1998,
Behavioral and Brain Sciences.
[4]
G. C. Quarton,et al.
The neurosciences : a study program
,
1967
.
[5]
Peter Cariani,et al.
Towards an Evolutionary Semiotics: The Emergence of New Sign-Functions in Organisms and Devices
,
1998
.
[6]
Charles E. Taylor,et al.
Artificial Life II
,
1991
.
[7]
E. Capaldi,et al.
The organization of behavior.
,
1992,
Journal of applied behavior analysis.
[8]
B. Delgutte,et al.
Neural correlates of the pitch of complex tones. I. Pitch and pitch salience.
,
1996,
Journal of neurophysiology.
[9]
Lutz Prechelt,et al.
Exploiting Domain-Specific Properties: Compiling Parallel Dynamic Neural Network Algorithms into Efficient Code
,
1999,
IEEE Trans. Parallel Distributed Syst..
[10]
W. T. Powers.
Behavior, the control of perception
,
1973
.
[11]
E. Gibson.
Principles of Perceptual Learning and Development
,
1969
.