by Jurgen Rahmel, Thomas Villmann
http://kbibmp3.ub.uni-kl.de/Preprint_Informatik/PS/lsa-96-01e.ps.gz
Add To MetaCart
Abstract:
In this report, we first propose a dichotomy of topology preserving network models based on the degree to which the structure of a network is determined by the given task. We then look closer at one of those groups and investigate the information that is contained in the graph structure of a topology preserving neural network. The task we have in mind is the usage of the network's topology for the retrieval of nearest neighbors of a neuron or a query, as it is of importance, e.g., in medical diagnosis systems. In general considerations, we propose certain properties of the structure and formulate the respective expectable results of network interpretation. From the results we conclude that both topology preservation as well as neuron distribution are highly influential for the network semantics. After a short survey on hierarchical models for data analysis, we propose a new network model that fits both needs. This so called SplitNet model dynamically constructs a hierarchically structured network that provides interpretability by neuron distribution, network topology and hierarchy of the network layers. We present empirical results for this new model and demonstrate its application in the medical domain of nerve lesion diagnosis. Further, we explain a view how the interpretation of the hierarchy in models like SplitNet can be understood in the context of integration of symbolic and connectionist learning. 1
Citations
|
3215
|
C4.5: Programs for machine learning
– Quinlan
- 1993
|
|
2489
|
Induction of Decision Trees
– Quinlan
- 1986
|
|
149
|
Pattern Recognition Principles
– Tou, Gonzalez
- 1974
|
|
138
|
Graph-Theoretical Methods for Detecting and Describing Gestalt Clusters
– Zahn
- 1971
|
|
101
|
Asymptotic quantization error of continuous signals and the quantization dimension
– Zador
- 1982
|
|
57
|
On the stationary state of Kohonen's self-organizing sensory mapping
– Ritter, Schulten
- 1986
|
|
53
|
Perceptron trees - a case study in hybrid concept representation
– Utgoff
- 1988
|
|
25
|
von der Malsburg, How patterned neural connections can be set up by self-organization
– Willshaw, C
- 1976
|
|
15
|
Vector quantization of image based upon the Kohonen self-organizing feature maps
– Nasrabadi, Feng
- 1988
|
|
8
|
Topology preservation in self-organizing feature maps: Exact definition and measurement
– Villmann, Herrmann, et al.
- 1994
|
|
6
|
A new quantitative measure of topology preservation in Kohonen's feature maps
– Villmann, Martinetz
- 1994
|
|
5
|
Neural Networks
– Martinetz, Schulten
- 1992
|
|
4
|
On the Role of Topology for Neural Network Interpretation
– Rahmel
- 1996
|
|
3
|
SplitNet: Learning of Hierarchical Kohonen Chains
– Rahmel
- 1996
|
|
2
|
Similarity-based self-organized clustering
– Rahmel
- 1995
|
|
1
|
SplitNet: A Dynamic Hierarchical Network Model
– Rahmel
- 1996
|
|
1
|
A Tree-Structured Approach to Medical Diagnosis Tasks
– Rahmel, Hahn
- 1996
|
