An incremental network model is introduced which is able to learn the important topological relations in a given set of input vectors by means of a simple Hebb-like learning rule. In contrast to previous approaches like the "neural gas" method of Martinetz and Schulten (1991, 1994), this model has no parameters which change over time and is able to continue learning, adding units and connections, until a performance criterion has been met. Applications of the model include vector quantization, clustering, and interpolation. 1 INTRODUCTION In unsupervised learning settings only input data is available but no information on the desired output. What can the goal of learning be in this situation? One possible objective is dimensionality reduction: finding a low-dimensional subspace of the input vector space containing most or all of the input data. Linear subspaces with this property can be computed directly by principal component analysis or iteratively with a number of network models (S...

... This article illustrates the potential learning capabilities of purely local learning and offers an interesting and powerful approach to learning with receptive fields.

Abstract not found

Recognition of human faces using a gallery of still or video images and a probe set of videos is systematically investigated using a probabilistic framework. In still-to-video recognition, where the gallery consists of still images, a time series state space model is proposed to fuse temporal information in a probe video, which simultaneously characterizes the kinematics and identity using a motion vector and an identity variable, respectively. The joint posterior distribution of the motion vector and the identity variable is estimated at each time instant and then propagated to the next time instant. Marginalization over the motion vector yields a robust estimate of the posterior distribution of the identity variable. A computationally ecient sequential importance sampling algorithm is developed to provide a numerical solution to the model. Theoretical derivations under weak assumptions demonstrate that, due to the propagation of the identity variable over time, a degeneracy in the posterior probability of the identity variable is exploited to give improved recognition.

Any agent in the real world has to be able to make distinctions between different types of objects, i.e. it must have the competence of categorization. In mobile agents, there is a large variation in proximal sensory stimulation originating from the same object. Therefore, categorization behavior is hard to achieve, and the successes in the past in solving this problem, have been limited. In this paper it is proposed that the problem of categorization in the real world is significantly simplified if it is viewed as one of sensory-motor coordination, rather than one of information processing happening "on the input side". A series of models is presented to illustrate the approach. It is concluded that we should consider replacing the metaphor of information processing for intelligent systems by the one of sensory-motor coordination. But the principle of sensory-motor coordination is more than a metaphor. It offers concrete mechanisms for putting agents to work in the real world. These i...

. We present a novel self-organizing network which is generated by a growth process. The application range of the model is the same as for Kohonen's feature map: generation of topologypreserving and dimensionality-reducing mappings, e.g., for the purpose of data visualization. The network structure is a rectangular grid which, however, increases its size during self-organization. By inserting complete rows or columns of units the grid may adapt its height/width ratio to the given pattern distribution. Both the neighborhood range used to co-adapt units in the vicinity of the winning unit and the adaptation strength are constant during the growth phase. This makes it possible to let the network grow until an application-specific performance criterion is fulfilled or until a desired network size is reached. A final approximation phase with decaying adaptation strength fine-tunes the network. 1 Introduction The self-organizing feature map [1] is a widely used method for generating topolog...

In this survey paper, we review the constructive algorithms for structure learning in feedforward neural networks for regression problems. The basic idea is to start with a small network, then add hidden units and weights incrementally until a satisfactory solution is found. By formulating the whole problem as a state space search, we first describe the general issues in constructive algorithms, with special emphasis on the search strategy. A taxonomy, based on the differences in the state transition mapping, the training algorithm and the network architecture, is then presented. Keywords--- Constructive algorithm, structure learning, state space search, dynamic node creation, projection pursuit regression, cascade-correlation, resource-allocating network, group method of data handling. I. Introduction A. Problems with Fixed Size Networks I N recent years, many neural network models have been proposed for pattern classification, function approximation and regression problems. Among...

Self-organizing maps are a prominent unsupervised neural network model providing cluster analysis of highdimensional input data. However, in spite of enhanced visualization techniques for self-organizing maps, interpreting a trained map proves to be difficult because the features responsible for a specific cluster assignment are not evident from the resulting map representation. In this paper we present our LabelSOM approach for automatically labeling a trained selforganizing map with the features of the input data that are the most relevant ones for the assignment of a set of input data to a particular cluster. The resulting labeled map allows the user to understand the structure and the information available in the map and the reason for a specific map organization, especially when only little prior information on the data set and its characteristics is available. We demonstrate the applicability of the LabelSOM method in the field of data mining providing an example from real world...

Clustering algorithms aim at modelling fuzzy (i.e., ambiguous) unlabeled patterns efficiently. Our goal is to propose a theoretical framework where clustering systems can be compared on the basis of their learning strategies. In the first part of this work, the following issues are reviewed: relative (probabilistic) and absolute (possibilistic) fuzzy membership functions and their relationships to the Bayes rule, batch and on-line learning, growing and pruning networks, modular network architectures, topologically perfect mapping, ecological nets and neuro-fuzziness. From this discussion an equivalence between the concepts of fuzzy clustering and soft competitive learning in clustering algorithms is proposed as a unifying framework in the comparison of clustering systems. Moreover, a set of functional attributes is selected for use as dictionary entries in our comparison. In the second part of this paper, five clustering algorithms taken from the literature are reviewed and compared on...

Document classification is one of the central issues in information retrieval research. The aim is to uncover similarities between text documents. In other words, classification techniques are used to gain insight in the structure of the various data items contained in the text archive. In this paper we show the results from using a hierarchy of self-organizing maps to perform the text classification task. Each of the individual self-organizing maps is trained independently and gets specialized to a subset of the input data. As a consequence, the choice of this particular artificial neural network model enables the true establishment of a document taxonomy. The benefit of this approach is a straightforward representation of document similarities combined with dramatically reduced training time. In particular, the hierarchical representation of document collections is appealing because it is the underlying organizational principle in use by librarians providing the necessary familiarity...