Effectiveness of document representation for classification

Conventionally, document classification researches focus on improving the learning capabilities of classifiers. Nevertheless, according to our observation, the effectiveness of classification is limited by the suitability of document representation. Intuitively, the more features that are used in representation, the more comprehensive that documents are represented. However, if a representation contains too many irrelevant features, the classifier would suffer from not only the curse of high dimensionality, but also overfitting. To address this problem of suitableness of document representations, we present a classifier-independent approach to measure the effectiveness of document representations. Our approach utilises a labelled document corpus to estimate the distribution of documents in the feature space. By looking through documents in this way, we can clearly identify the contributions made by different features toward the document classification. Some experiments have been performed to show how the effectiveness is evaluated. Our approach can be used as a tool to assist feature selection, dimensionality reduction and document classification.

Representation and learning of visual information for pose recognition

Recovering position from sensor information is an important problem in mobile robotics, known as localisation. Localisation requires a map or some other description of the environment to provide the robot with a context to interpret sensor data. The mobile robot system under discussion is using an artificial neural representation of position. Building a geometrical map of the environment with a single camera and artificial neural networks is difficult. Instead it would be simpler to learn position as a function of the visual input. Usually when learning images, an intermediate representation is employed. An appropriate starting point for biologically plausible image representation is the complex cells of the visual cortex, which have invariance properties that appear useful for localisation. The effectiveness for localisation of two different complex cell models are evaluated. Finally the ability of a simple neural network with single shot learning to recognise these representations and localise a robot is examined.

A longitudinal investigation of self-other discrimination and the emergence of mirror self-recognition

The aim in this study was to investigate the association between infants' developing interest in their self-image and the onset of mirror self-recognition (MSR). A longitudinal study was conducted with 98 infants who were seen at intervals of 3 months from 9-24 months of age. At each session, the infants were administered a preferential-looking test whereby they were presented with a video image of themselves alongside a video image of a same-aged peer in two conditions, unmarked and marked. From the 12-month session onwards, the infants were also administered a version of the standard mark test of MSR. The infants showed a significant preference for looking at images of themselves in both conditions coincident with the onset of MSR. This result indicates that developing an interest in the self-image is an important component in the development of MSR. (C) 2003 Elsevier Science Inc. All rights reserved.

Using the EM Algorithm to Train Neural Networks: Misconceptions and a New Algorithm for Multiclass Classification

The expectation-maximization (EM) algorithm has been of considerable interest in recent years as the basis for various algorithms in application areas of neural networks such as pattern recognition. However, there exists some misconceptions concerning its application to neural networks. In this paper, we clarify these misconceptions and consider how the EM algorithm can be adopted to train multilayer perceptron (MLP) and mixture of experts (ME) networks in applications to multiclass classification. We identify some situations where the application of the EM algorithm to train MLP networks may be of limited value and discuss some ways of handling the difficulties. For ME networks, it is reported in the literature that networks trained by the EM algorithm using iteratively reweighted least squares (IRLS) algorithm in the inner loop of the M-step, often performed poorly in multiclass classification. However, we found that the convergence of the IRLS algorithm is stable and that the log likelihood is monotonic increasing when a learning rate smaller than one is adopted. Also, we propose the use of an expectation-conditional maximization (ECM) algorithm to train ME networks. Its performance is demonstrated to be superior to the IRLS algorithm on some simulated and real data sets.

Pretend play, mirror self-recognition and imitation: A longitudinal investigation through the second year

The aim in the current study was to investigate the emergence of pretend play, mirror self-recognition, synchronic imitation and deferred imitation in normally developing human infants. A longitudinal study was conducted with 98 infants seen at three-monthly intervals from 12 through to 24 months of age. At each session the infants were tested on a range of tasks assessing the four target skills. Deferred imitation was found to emerge prior to synchronic imitation, pretend play and mirror self-recognition. In contrast, the latter three skills emerged between 18 and 21 months and followed similar developmental trajectories. Deferred imitation was found to hold a prerequisite relation with these three skills. Synchronic imitation, pretend play and mirror self-recognition were not closely associated and no prerequisite relations were found between these skills. These findings are discussed in the context of current theories regarding the development of pretend play, mirror self-recognition, synchronic imitation and deferred imitation in the second year. (C) 2004 Elsevier Inc. All rights reserved.

Bioactivity of Fusarium oxysporum f. sp glycines and Sclerotium rolfsii filtrates on egg hatching, survival and infectivity of juveniles of Meloidogyne incognita race 2

The effects of culture filtrates of Fusarium oxysporum and Sclerotium rolfsii on egg hatching and juvenile survival of Meloidogyne incognita in vitro and impact of these filtrates on infectivity of M. incognita were investigated on soybean seedlings. Five- and 10-day-old filtrates of F. oxysporum caused 65 and 54% egg-hatching inhibition, while that of S. rolfsii caused 61 and 49% inhibition, respectively. Juveniles of M. incognita died within 6 days when incubated in 5-day-old filtrate of F. oxysporum, while the similar filtrate of S. rolfsii caused 100% juvenile mortality on the fifth day. Filtrates reduced root galling, egg population, number of adult females in soybean plants at harvest and also soil population. Culture filtrates could be used as source of biological nematicides.