Abstract | ||
---|---|---|
It is often that the learned neural networks end with different decision boundaries under the variations of training data, learning algorithms, architectures, and initial random weights. Such variations are helpful in designing neural network ensembles, but are harmful for making unstable performances, i.e., large variances among different learnings. This paper discusses how to reduce such variances for learned neural networks by letting them re-learn on those data points on which they disagrees with each other. Experimental results have been conducted on four real world applications to explain how and when such re-learning works. |
Year | DOI | Venue |
---|---|---|
2008 | 10.1109/IJCNN.2008.4634054 | 2008 IEEE INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, VOLS 1-8 |
Keywords | Field | DocType |
decision boundary,artificial neural networks,heart,neural networks,diabetes,cancer,neural nets,stability,training data,neural network,testing | Data point,Training set,Computer science,Artificial intelligence,Artificial neural network,Decision boundary,Machine learning | Conference |
ISSN | Citations | PageRank |
2161-4393 | 0 | 0.34 |
References | Authors | |
4 | 1 |