. A market-based algorithm is presented which autonomously apportions complex tasks to multiple cooperating agents giving each agent the motivation of improving performance of the whole system. A specific model, called "The Hayek Machine" is proposed and tested on a simulated Blocks World (BW) planning problem. Hayek learns to solve more complex BW problems than any previous learning algorithm. Given intermediate reward and simple features, it has learned to efficiently solve arbitrary BW problems. The Hayek Machine can also be seen as a model of evolutionary economics. 1 Introduction There is a growing consensus that one should attempt to understand the mind as arising from the interaction of many modules or agents which are much simpler than the whole, yet complex compared to individual neurons. This view is separately expressed by workers in a wide variety of disciplines, each discipline offering independent empirical evidence, c.f. the literatures of evolutionary psychology[14], ...
|
2044
|
Learning internal representations by error propagation
– Rumelhart, G, et al.
- 1986
|
|
968
|
Learning from delayed rewards
– Watkins
- 1989
|
|
940
|
Unified theories of cognition
– Newell
- 1990
|
|
931
|
Learning to predict by the methods of temporal differences
– Sutton
- 1988
|
|
713
|
Genetic Programming
– Koza
- 1992
|
|
567
|
A Framework for Representing Knowledge
– Minsky
- 1975
|
|
559
|
An Evolutionary Theory of Economic Change
– Nelson, Winter
- 1982
|
|
406
|
Learning to act using real-time dynamic programming
– Barto, Bradtke, et al.
- 1995
|
|
376
|
The tragedy of the commons
– Hardin
- 1968
|
|
353
|
Systematic nonlinear planning
– McAllester, Rosenblitt
- 1991
|
|
320
|
The society of mind. Simon and
– Minsky
- 1986
|
|
284
|
The nature of the firm
– Coase
- 1937
|
|
261
|
Understanding Natural Language
– Winograd
|
|
251
|
an Approach to the Synthesis of Life
– Ray
- 1992
|
|
243
|
A market-oriented programming environment and its application to distributed multicommodity ow problems
– Wellman
- 1993
|
|
223
|
Improving elevator performance using reinforcement learning
– Crites, Barto
- 1996
|
|
183
|
Models of bounded rationality
– Simon
- 1982
|
|
87
|
Pandemonium: A Paradigm for Learning
– Selfridge
- 1959
|
|
81
|
Cognitive adaptations for social exchange
– Cosmides, Tooby
- 1992
|
|
79
|
Escaping brittleness: The possibilities of general purpose learning algorithms applied to parallel rule-based systems
– Holland
- 1986
|
|
76
|
The Ecology of Computation
– Huberman
- 1988
|
|
69
|
Circuits of the Mind
– Valiant
- 1994
|
|
63
|
PRODIGY4.0: The Manual and Tutorial
– Carbonell
- 1992
|
|
55
|
Artificial economic life: a simple model of a stockmarket. Physica D: Nonlinear Phenomena
– Palmer, Arthur, et al.
- 1994
|
|
51
|
Learning to reason
– Khardon
- 1997
|
|
46
|
On the complexity of domain-independent planning
– Erol, Nau, et al.
- 1992
|
|
42
|
The role of heuristics in learning by discovery: Three case studies
– Lenat
|
|
40
|
T.: High-performance job-shop scheduling with a timedelay TD (λ) network
– Zhang, Dietterich
- 1996
|
|
39
|
A critical review of classifier systems
– Wilson
- 1989
|
|
30
|
The Economy as an Evolving Complex System: The
– Anderson, Arrow, et al.
- 1988
|
|
30
|
Multi-strategy learning of search control for partial-order planning
– Estlin, Mooney
- 1996
|
|
25
|
Toward a Model of Mind as a Laissez-Faire Economy of
– Baum
- 1996
|
|
21
|
The Working Brain - An Introduction in Neuropsychology. Basic Books
– Luria
- 1973
|
|
19
|
Learning to perceive and act
– Whitehead, Ballard
- 1991
|
|
14
|
On genetic algorithms
– Baum, Boneh, et al.
- 1995
|
|
13
|
Hill climbing beats genetic search on a boolean circuit synthesis problem of koza's
– Lang
- 1995
|
|
11
|
Incremental Learning of Evaluation Functions for Absorbing Markov Chains: New Methods and Theorems” preprint
– Gurvits, Lin, et al.
- 1994
|
|
11
|
Representational Difficulties with Classifier Systems
– Schuurmans, Schaeffer
- 1989
|
|
9
|
Adaptation in dynamic environments through a minimal probability of exploration
– Venturini
- 1994
|
|
8
|
Implementing semantic network structures using the classifier system
– Forrest
- 1985
|
|
8
|
Rationality
– Valiant
- 1995
|
|
4
|
Steps towards Artificial Intelligence," Computers and Thought
– Minsky
- 1963
|
|
3
|
The SNLP planner implementation. Contact bugsnlp @cs.washington.edu
– Barrett, Weld
- 1990
|
|
2
|
Using temporal logic to control search in planning, unpublished document available from http://logos.uwaterloo.ca/tlplan/tlplan.html
– Bacchus, Kabanza
- 1995
|
|
2
|
Markets and Computation: Agoric Open Systems," The Ecology of Computation
– Miller, Drexler
- 1988
|
|
2
|
Roadkill on the information highway
– Myhrvold
- 1994
|
|
2
|
Practical issues in temporal difference learing
– Tesauro
- 1992
|
|
1
|
Esben Sloth, Evolutionary Economics: Post-Schumpeterian Contributions
– Andersen
- 1996
|
|
1
|
Economic Metalearning, submitted for publication
– Baum, Durdanovic
- 1997
|
|
1
|
issue of "Reason
– Coase
- 1997
|
