Andrew R. McIntyre

Symbiotic coevolutionary genetic programming: a benchmarking study under large attribute spaces

Classification under large attribute spaces represents a dual learning problem in which attribute subspaces need to be identified at the same time as the classifier design is established. Embedded as opposed to filter or wrapper methodologies address both tasks simultaneously. The motivation for this work stems from the observation that team based approaches to Genetic Programming (GP) have the potential to design multiple classifiers per class—each with a potentially unique attribute subspace—without recourse to filter or wrapper style preprocessing steps. Specifically, competitive coevolution provides the basis for scaling the algorithm to data sets with large instance counts; whereas cooperative coevolution provides a framework for problem decomposition under a bid-based model for establishing program context. Symbiosis is used to separate the tasks of team/ensemble composition from the design of specific team members. Team composition is specified in terms of a combinatorial search performed by a Genetic Algorithm (GA); whereas the properties of individual team members and therefore subspace identification is established under an independent GP population. Teaming implies that the members of the resulting ensemble of classifiers should have explicitly non-overlapping behaviour. Performance evaluation is conducted over data sets taken from the UCI repository with 649–102,660 attributes and 2–10 classes. The resulting teams identify attribute spaces 1–4 orders of magnitude smaller than under the original data set. Moreover, team members generally consist of less than 10 instructions; thus, small attribute subspaces are not being traded for opaque models.

Classification as clustering: A pareto cooperative-competitive gp approach

Intuitively population based algorithms such as genetic programming provide a natural environment for supporting solutions that learn to decompose the overall task between multiple individuals, or a team. This work presents a framework for evolving teams without recourse to prespecifying the number of cooperating individuals. To do so, each individual evolves a mapping to a distribution of outcomes that, following clustering, establishes the parameterization of a (Gaussian) local membership function. This gives individuals the opportunity to represent subsets of tasks, where the overall task is that of classification under the supervised learning domain. Thus, rather than each team member representing an entire class, individuals are free to identify unique subsets of the overall classification task. The framework is supported by techniques from evolutionary multiobjective optimization (EMO) and Pareto competitive coevolution. EMO establishes the basis for encouraging individuals to provide accurate yet nonoverlaping behaviors; whereas competitive coevolution provides the mechanism for scaling to potentially large unbalanced datasets. Benchmarking is performed against recent examples of nonlinear SVM classifiers over 12 UCI datasets with between 150 and 200,000 training instances. Solutions from the proposed coevolutionary multiobjective GP framework appear to provide a good balance between classification performance and model complexity, especially as the dataset instance count increases.

Automated optic nerve analysis for diagnostic support in glaucoma

The availability of modern imaging techniques such as confocal scanning laser tomography (CSLT) for capturing high-quality optic nerve images offer the potential for developing automatic and objective methods for supporting clinical decision-making in glaucoma. We present a hybrid approach that features the analysis of CSLT images using moment methods to derive abstract image defining features, and the use of these features to train classifiers for automatically distinguishing CSLT images of healthy and diseased optic nerves. As a first step, in this paper, we present investigations in feature subset selection methods for reducing the relatively large input space produced by the moment methods. Our results demonstrate that our methods discriminate between healthy and glaucomatous optic nerves based on shape information automatically derived from CSLT tomography images.

Evolving GP Classifiers for Streaming Data Tasks with Concept Change and Label Budgets: A Benchmarking Study

Streaming data classification requires that several additional challenges are addressed that are not typically encountered in offline supervised learning formulations. Specifically, access to data at any training generation is limited to a small subset of the data, and the data itself is potentially generated by a non-stationary process. Moreover, there is a cost to requesting labels, thus a label budget is enforced. Finally, an anytime classification requirement implies that it must be possible to identify a ‘champion’ classifier for predicting labels as the stream progresses. In this work, we propose a general framework for deploying genetic programming (GP) to streaming data classification under these constraints. The framework consists of a sampling policy and an archiving policy that enforce criteria for selecting data to appear in a data subset. Only the exemplars of the data subset are labeled, and it is the content of the data subset that training epochs are performed against. Specific recommendations include support for GP task decomposition/modularity and making additional training epochs per data subset. Both recommendations make significant improvements to the baseline performance of GP under streaming data with label budgets. Benchmarking issues addressed include the identification of datasets and performance measures.

Toward glaucoma classification with moment methods

This paper presents a series of experiments testing the feasibility of employing image-processing techniques for the feature extraction stage in the implementation of a basic optic nerve image classifier. Such a scheme completely removes the need for manually identifying the edge of the optic nerve. In this work, Zernike moments are extracted from Confocal Scanning Laser Tomography images of optic discs for the purposes of classifying the disc as healthy or damaged using a linear discriminant function derived from a linear perceptron. Our preliminary results, when compared with the performance of conventional feature sets, demonstrate the appropriateness of this approach.

Exploring content-based image indexing techniques in the compressed domain

A method is proposed for performing clustering and offline indexing on the signal representations of compressed JPEG images. The current system clusters on discrete cosine transform (DCT) blocks of JPEG images using the potential function clustering algorithm, storing indices of varying length for a posteriori comparison and processing of query images. Results presented indicate the appropriateness of using clusters derived from JPEG DCT blocks for content-based indexing. In particular, we are able to provide image summaries based on index features defined from a reference texture database, which significantly speeds the search process.

On the application of GP to streaming data classification tasks with label budgets

A framework is introduced for applying GP to streaming data classification tasks under label budgets. This is a fundamental requirement if GP is going to adapt to the challenge of streaming data environments. The framework proposes three elements: a sampling policy, a data subset and a data archiving policy. The sampling policy establishes on what basis data is sampled from the stream, and therefore when label information is requested. The data subset is used to define what GP individuals evolve against. The composition of such a subset is a mixture of data forwarded under the sampling policy and historical data identified through the data archiving policy. The combination of sampling policy and the data subset achieve a decoupling between the rate at which the stream passes and the rate at which evolution commences. Benchmarking is performed on two artificial data sets with specific forms of sudden shift and gradual drift as well as a well known real-world data set.

On Multi-class Classification by Way of Niching

In recent literature, the niche enabling effects of crowding and the sharing algorithms have been systematically investigated in the context of Genetic Algorithms and are now established evolutionary methods for identifying optima in multi-modal problem domains. In this work, the niching metaphor is methodically explored in the context of a simultaneous multi-population GP classifier in order to investigate which (if any) properties of traditional sharing and crowding algorithms may be portable in arriving at a naturally motivated niching GP. For this study, the niching mechanisms are implemented in Grammatical Evolution to provide multi-category solutions from the same population in the same trial. Each member of the population belongs to a different niche in the GE search space corresponding to the data classes. The set of best individuals from each niche are combined hierarchically and used for multi-class classification on the familiar multi-class UCI data sets of Iris and Wine. A distinct preference for Sharing as opposed to Crowding is demonstrated with respect to population diversity during evolution and niche classification accuracy.

Pareto Cooperative-Competitive Genetic Programming: A Classification Benchmarking Study

A model for problem decomposition in Genetic Programming based classication is proposed consisting of four basic components: competitive coevolution, local Gaussian wrapper operators, evolutionary multiobjective (EMO) tness evaluation, and an explicitly cooperative objective. The framework specically emphasizes the relations between different components of the model. Thus, both the local wrapper operator and cooperative objective components work together to establish exemplar subsets against which performance is evaluated and the decomposition of the problem domain is achieved. Moreover, the cost of estimating tness over multiple objectives is mitigated by the ability to associate specic subsets of exemplars with each classier.

Label free change detection on streaming data with cooperative multi-objective genetic programming

Classification under streaming data conditions requires that the machine learning (ML) approach operate interactively with the stream content. Thus, given some initial ML classification capability, it is not possible to assume that stream content will be stationary. It is therefore necessary to first detect when the stream content changes. Only after detecting a change, can classifier retraining be triggered. Current methods for change detection tend to assume an entropy filter approach, where class labels are necessary. In practice, labeling the stream would be extremely expensive. This work proposes an approach in which the behaviour of GP individuals is used to detect change without the use of labels. Only after detecting a change is label information requested. Benchmarking under a computer network traffic analysis scenario demonstrates that the proposed approach performs at least as well as the filter method, while retaining the advantage of requiring no labels.