Abstract:
1 . Sensors represent a crucial link between the evolutionary forces shaping a species' relationship with its environment, and the individual's cognitive abilities to behave and learn. We report on experiments using a new class of "latent energy environments" (LEE) models to define environments of carefully controlled complexity which allow us to state bounds for random and optimal behaviors that are independent of strategies for achieving the behaviors. Using LEE's analytic basis for defining environments, we then use neural networks (NNets) to model individuals and a steadystate genetic algorithm to model an evolutionary process shaping the NNets, in particular their sensors. Our experiments consider two types of "contact" and "ambient" sensors, and variants where the NNets are not allowed to learn, learn via error correction from internal prediction, and via reinforcement learning. We find that predictive learning, even when using a larger repertoire of the more sophisticated ambi...
Citations
|
2044
|
Learning internal representations by error propagation
– Rumelhart, G, et al.
- 1986
|
|
178
|
Interactions between learning and evolution
– Ackley, Littman
- 1991
|
|
113
|
Learning and Evolution in Neural Networks
– Nolfi, Elman, et al.
- 1994
|
|
105
|
The spandrels of San Marco and the panglossian paradigm: A critique of the adaptationist programme
– Gould, Lewontin
- 1979
|
|
68
|
Econets: Neural Networks that Learn in a Environment
– Parisi, Cecconi
- 1990
|
|
57
|
Pattern recognizing stochastic learning automata
– Barto
- 1985
|
|
48
|
Antfarm: Towards simulated evolution
– Collins, Jefferson
- 1991
|
|
47
|
The evolution of life histories
– Stearns
- 1992
|
|
28
|
Genetic Reinforcement Learning With Multilayer Neural Networks
– Whitley, Dominic, et al.
- 1991
|
|
25
|
Generation gaps revisited
– Jong, Sarma
- 1993
|
|
22
|
Adaptation in Natural and Artificial Systems, second edition
– Holland
- 1992
|
|
14
|
Analysis of Evolved Sensory-Motor Controllers
– Cliff, Husbands, et al.
- 1992
|
|
6
|
Changing Latent Energy Environments: A Case for the Evolution of Plasticity
– Menczer
- 1994
|
|
6
|
Recombination and unsupervised learning: effects of crossover in the genetic optimization of neural networks
– Menczer, Parisi
- 1992
|
|
4
|
Genetic Algorithms and Artificial Life”. Santa Fe Institute Working Paper 93-11-072
– Mitchell, Forrest
- 1994
|
|
2
|
Introduction to the ECHO model. Santa Fe Institute. Working paper 93-12-074 Complex Environments to Complex Behaviors Filippo Menczer and Richard K. Belew Adaptive Behavior Volume 4, Number 3/4, 317-363 Preprint Page 40 Kaelbling, L.P
– Jones, Mitchell
- 1993
|
|
2
|
Evolutionary stable and unstable strategies in neural networks
– Miglino, Parisi
- 1991
|
|
1
|
Generalization and Scaling
– Ackley, Littman
- 1990
|
|
1
|
LEE: A Tool for Artificial Life Simulations
– Menczer, Belew
- 1993
|
|
1
|
Latent Energy Environments. Plastic Individuals in Evolving Populations
– Menczer, Belew
- 1994
|
|
1
|
Adequate Locomotion Strategies for an Abstract Environment -- A Relational Approach
– Rssler
- 1974
|
