Learning anticipatory behaviour using a delayed action classifier system

A learning classifier system is one of the methods for applying a genetic-based approach to machine learning applications. An enhanced version of the system that employs the Bucket-brigade algorithm to reward individuals in a chain of co-operating rules is implemented and assigned the task of learning rules for control robotics behaviors. Illustrates the approach with the example of kicking a moving Ball into a goal (KMB). The KMB (kick a Moving Ball) System built of two-classifier subsystems work together, each classifier system learns a simple behavior, first classifier system, learn simulated robot chase behavior i.e. learn robot to move single step toward moving ball, second classifier system, learn the simulated robot approach behavior i.e. learn robot to kick the ball toward fixed goal, the system as a whole has as its learning goal the coordinate of behaviors. This work examine the performance of the simple classifier system (SCS) on the (KMB) problem, perform two SCS simulations one without the genetic algorithm enabled (GA) and one with the genetic algorithm enabled (GA) results using a classifier system with genetic algorithm show improvement over one without, and furthermore the level of performance has been high enough to rival human accuracy. Also the results without the genetic algorithm show that the apportionment of credit algorithm adjust the strength values of the rules.

In modern complex environment with continuous updating information, knowledge acquisition has become a bottleneck for the expert systems. A symbolic artificial intelligence (AI) system may possess reasoning knowledge with interpretability, but lack of learning capability, especially when environment status and dynamic trend change, and habits of input condition switch. Numerical artificial intelligence systems, on the other hand, possess relearning capability, but lack of interpretability, such as a bunch of hard intuitive and hard recognizing weights for representing knowledge. Therefore, integrating the advantages of the two types of systems can be an important issue in current and future research. Holland first proposed the classifiers system in 1975. The classifiers systems are genetic-algorithm based, incorporate the reinforcement learning, and constitute one of the theories of machine learning. It uses simple binary-variables (1, 0, #) for knowledge representation and expedites the system's responsive speed. Nevertheless, there is a limitation if it encountered with environmental uncertainty or fuzziness, continuous input variables and fuzzy feedback. Based on these demands described above, this paper proposes a fuzzy message requirement classifiers system (FMRCS), which integrates learning and reasoning. The FMRCS consists of two systems: the message requirement classifiers system (MRCS) and the fuzzy classifiers system (FCS). First, the MRCS simulates the internal reasoning process of human being and finds the key factors of decision problem under limited known messages in the problem domain. Then, according to the key factors created by the MRCS, the FCS learns the relationships between all factors concerned of the problem domain and various results of this problem.

The class imbalance problem has been said recently to hinder the performance of learning systems. In fact, many of them are designed with the assumption of well-balance datasets. However, it is very common to find higher presence of one of the classes in real classification problems. The aim of this paper is to make a preliminary analysis on the effect of the class imbalance problem in learning classifier systems. Particularly we focus our study on UCS, a supervised version of XCS classifier system. We analyze UCS's behavior on unbalanced datasets and find that UCS is sensitive to high levels of class imbalance. We study strategies for dealing with class imbalances, acting either at the sampling level or at the classifier system's level.

A Fuzzy Classifier System for evolutionary learning of robot behaviors

Classifier systems and genetic algorithms

Operating rules classification system of water supply reservoir based on Learning Classifier System

It has long been known that in some relatively simple reinforcement learning tasks traditional strength-based classifier systems will adapt poorly and show poor generalisation. In contrast, the more recent accuracy-based XCS, appears both to adapt and generalise well. In this work, we attribute the difference to what we call strong over general and fit over general rules. We begin by developing a taxonomy of rule types and considering the conditions under which they may occur. In order to do so an extreme simplification of the classifier system is made, which forces us toward qualitative rather than quantitative analysis. We begin with the basics, considering definitions for correct and incorrect actions, and then correct, incorrect, and overgeneral rules for both strength and accuracy-based fitness. The concept of strong overgeneral rules, which we claim are the Achilles' heel of strength-based classifier systems, are then analysed. It is shown that strong overgenerals depend on what we call biases in the reward function (or, in sequential tasks, the value function). We distinguish between strong and fit overgeneral rules, and show that although strong overgenerals are fit in a strength-based system called SB-XCS, they are not in XCS. Next we show how to design fit overgeneral rules for XCS (but not SB-XCS), by introducing biases in the variance of the reward function, and thus that each system has its own weakness. Finally, we give some consideration to the prevalence of reward and variance function bias, and note that non-trivial sequential tasks have highly biased value functions.

We present a complete design of agents for the RoboCup Rescue Simulation problem that uses an evolutionary reinforcement learning mechanism called XCS, a version of Holland’s Genetic Classifiers Systems, to decide the number of ambulances required to rescue a buried civilian. We also analyze the problems implied by the rescue simulation and present solutions for every identified sub-problem using multi-agent cooperation and coordination built over a subsumption architecture. Our agents’ classifier systems were trained in different disaster situations. Trained agents outperformed untrained agents and most participants of the 2004 RoboCup Rescue Simulation League competition. This system managed to extract general rules that could be applied on new disaster situations, with a computational cost of a reactive rule system.KeywordsAlive AgentDisaster SituationEvolutionary ReinforcementSystem Genetic ProgramAgent TeamThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

The authors examine aspects of machine learning by classifier systems that use genetic algorithms. In particular, adaptive image learning and classification are considered. Standard classifier systems are not well suited for seeking out multiple goals as is necessary in image learning and classification problems. To improve the performance of standard classifier systems for the image learning task, several modifications are suggested. The modifications result in a far better performance for classifier system on the ImageLearn domain. >

The agent computing paradigm is rapidly emerging as one of the powerful technology to deal with the uncertainty in dynamic environment. Recently, traditional learning classifier system are challenged by changes in the context. In this paper, an anticipatory agent based on the Anticipatory Learning Classifier System (ACS) for learning in changing environments is presented. This research aims to develop an agent learning architecture using anticipatory system that will enable intelligent agent to be able to detect environmental changes, adapt functionality at run-time to achieve goal. For achieving the intended target, an extension to the ACS framework called “Greedy Covering” to the ACS framework have been proposed. The novelty of the approach is in determining the changes in the environment and to generate optimal rules to adapt and reestablish the optimal policy to reach the goal state. The proposed algorithm is evaluated on several synthetic maze design and simulate a variety of changing environments. Experiment results indicate that up to 65% changes in an environment the ACS with the greedy covering can reestablish the optimal performance without increasing the number of classifiers.

In this paper, the use of genetic algorithms and classifiers systems in the graph coloring application was considered. A crossover derivative was developed to accommodate the graph coloring specifications. Also, a method for verifying and validating the new offspring coloring rules and regulations was considered. Through a series of experiments conducted on randomly generated graphs of a relatively small size, the importance of various parameters and their impact to the overall system performance is demonstrated and discussed. In conclusion, suggestions for future study regarding this topic are proposed and briefly discussed.

It is well known that standard learning classifier systems, when applied to many different domains, exhibit a number of problems: payoff oscillation, difficulty in regulating interplay between the reward system and the background genetic algorithm (GA), rule chains' instability, default hierarchies' instability, among others. ALECSYS is a parallel version of a standard learning classifier system (CS) and, as such, suffers from these same problems. In this paper we propose some innovative solutions to some of these problems. We introduce the following original features. Mutespec is a new genetic operator used to specialize potentially useful classifiers. Energy is a quantity introduced to measure global convergence to apply the genetic algorithm only when the system is close to a steady state. Dynamic adjustment of the classifiers set cardinality speeds up the performance phase of the algorithm. We present simulation results of experiments run in a simulated two-dimensional world in which a simple agent learns to follow a light source.

Combination of EEG and ECG for improved automatic neonatal seizure detection

We present ASDSP (application-specific digital signal processor) instructions and their hardware architecture for high-speed FFT. The proposed instructions calculate a butterfly within two cycles. The proposed architecture employs a data processing unit (DPU) supporting the instructions and an FFT address generation unit (FAGU) automatically calculating the butterfly input and output data addresses. The proposed DPU has a smaller area than commercial DSP chips. Moreover, the number of FFT computation cycles is reduced by the proposed FAGU. The architecture has been modeled by the VHDL. We have used the UMC 0.25/spl square/standard cell library for logic synthesis. Performance comparisons show that the number of execution cycles is reduced over 10% and the size of the DPU decreases about 30% compared with Carmel DSP.

System-level designers typically rely on white-box detailed descriptions of embedded systems in order to perform their design choices and optimisations. The white-box approach assumes that the complexity of the system can be managed by decomposing the global system into a set of simple and well-described components. This assumption is generally not true when non-linear interactions happen between the different subsystems. This paper explores, as an alternative, a black-box modelling methodology based on intelligent data analysis in order to support system-level designers in characterising the performance of embedded applications. The rationale behind this approach is that nowadays the designer can easily store large amount of data related to the system, and consequently discover relevant information by analysing them in a black-box fashion. As a case study, we address the performance modelling of software applications running on an embedded microprocessor. We introduce a data analysis method which, on the basis of a high-level characterisation of the software functionality and the hardware architecture, is able to predict the number of execution cycles on a embedded processor. We propose the adoption of a local learning technique (lazy learning), which proved already to be effective in previous works, to model the unknown input/output relation between the hardware/software parameters of the application and the number of execution cycles. Experiments with standard computational code (sorting, mathematical computation) and with a specific streaming algorithm for MPEG variable length decoding are presented to support this claim.