Online Learning with Random Representations

@inproceedings{Sutton1993OnlineLW,
  title={Online Learning with Random Representations},
  author={R. Sutton and S. Whitehead},
  booktitle={ICML},
  year={1993}
}
We consider the requirements of online learning|learning which must be done incrementally and in realtime, with the results of learning available soon after each new example is acquired. [...] Key Result Our results suggest that randomness has a useful role to play in online supervised learning and constructive induction. 1. Online Learning Applications of supervised learning can be divided into two types: online and oine.Expand
Representation Search through Generate and Test
  • A. Mahmood, R. Sutton
  • Computer Science
  • AAAI Workshop: Learning Rich Representations from Low-Level Sensors
  • 2013
TLDR
This work views the problem of representation learning as one of learning features such that performance of the underlying base system continually improves, and develops new representation-search methods and shows that the generate-and-test approach can be utilized in a simple and effective way for learning representations. Expand
Online Extreme Evolutionary Learning Machines
TLDR
It is demonstrated that a Darwinian neurodynamic approach of feature replication can improve performance beyond selection alone, and may offer a path towards effective learning of predictive models in robotic agents. Expand
Position Paper: Representation Search through Generate and Test
TLDR
This work views the problem of representation learning as one of learning features such that performance of the underlying base system continually improves, and develops new representation-search methods and shows that the generate-and-test approach can be utilized in a simple and effective way for continually improving representations. Expand
The Interplay of Search and Gradient Descent in Semi-stationary Learning Problems
We explore the interplay of generate-and-test and gradient-descent techniques for solving online supervised learning problems. The task in supervised learning is to learn a function using samples ofExpand
Generalization in Reinforcement Learning: Successful Examples Using Sparse Coarse Coding
TLDR
It is concluded that reinforcement learning can work robustly in conjunction with function approximators, and that there is little justification at present for avoiding the case of general λ. Expand
Online Discovery of Feature Dependencies
Online representational expansion techniques have improved the learning speed of existing reinforcement learning (RL) algorithms in low dimensional domains, yet existing online expansion methods doExpand
Online Representation Search and Its Interactions with Unsupervised Learning
We consider the problem of finding good hidden units, or features, for use in multilayer neural networks. Solution methods that generate candidate features, evaluate them, and retain the most usefulExpand
A survey of reinforcement learning in relational domains
TLDR
The aim is to give a complete survey of the available literature, of the underlying motivations and of the implications if the new methods for learning in large, relational and probabilistic environments. Expand
A Bayesian Approach to Finding Compact Representations for Reinforcement Learning
Feature-based function approximation methods have been applied to reinforcement learning to learn policies in a data-efficient way, even when the learner may not have visited all states duringExpand
A factorization perspective for learning representations in reinforcement learning
Reinforcement learning is a general formalism for sequential decision-making, with recent algorithm development focusing on function approximation to handle large state spaces and high-dimensional,Expand
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 56 REFERENCES
Fast Learning in Networks of Locally-Tuned Processing Units
We propose a network architecture which uses a single internal layer of locally-tuned processing units to learn both classification tasks and real-valued function approximations (Moody and DarkenExpand
Feature discovery by competitive learning
TLDR
This paper shows how a set of feature detectors which capture important aspects of the set of stimulus input patterns are discovered and how these feature detectors form the basis of a multilayer system that serves to learn categorizations of stimulus sets which are not linearly separable. Expand
Connectionist Learning Procedures
TLDR
These relatively simple, gradient-descent learning procedures work well for small tasks and the new challenge is to find ways of improving their convergence rate and their generalization abilities so that they can be applied to larger, more realistic tasks. Expand
Sparse Distributed Memory
TLDR
Pentti Kanerva's Sparse Distributed Memory presents a mathematically elegant theory of human long term memory that resembles the cortex of the cerebellum, and provides an overall perspective on neural systems. Expand
Practical Issues in Temporal Difference Learning
This paper examines whether temporal difference methods for training connectionist networks, such as Sutton's TD(λ) algorithm, can be successfully applied to complex real-world problems. A number ofExpand
Enhancing Transfer in Reinforcement Learning by Building Stochastic Models of Robot Actions
TLDR
Experiments on simulated and real robots show that useful action models can be learned from a 12 by 12 scrolling certainty grid sensor and on the simulator these models are sufficiently rich to enable significant transfer within and across two tasks, box pushing and wall following. Expand
Optimal Hidden Units for Two-Layer nonlinear Feedforward Neural Networks
  • T. Sanger
  • Mathematics, Computer Science
  • Int. J. Pattern Recognit. Artif. Intell.
  • 1991
TLDR
A definition is proposed for optimal nonlinear features, and a constructive method, which has an iterative implementation, is derived for finding them. Expand
Predicting the Future: Advantages of Semilocal Units
TLDR
It is postulate that an interesing behavior displayed by gaussian bar functions under gradient descent dynamics, which is called automatic connection pruning, is an important factor in the success of this representation. Expand
The self-organizing map
TLDR
The self-organizing map, an architecture suggested for artificial neural networks, is explained by presenting simulation experiments and practical applications, and an algorithm which order responses spatially is reviewed, focusing on best matching cell selection and adaptation of the weight vectors. Expand
An Evolutionary Pattern Recognition Network
A pattern recognition network with two types of adaptation has been investigated. The network output is a weighted sum of the outputs of elements which compute real functions of the discrete networkExpand
...
1
2
3
4
5
...