Domain transfer extreme learning machine method based on manifold regularization and norm regularization

Abstract

The invention discloses a domain transfer extreme learning machine method based manifold regularization and norm regularization. On the basis of a traditional extreme learning machine, the thought of semi-supervised learning and transfer learning is introduced, and a novel extreme learning machine model is built and consists of three parts: a manifold regularization term capable of excavating geometric distribution shapes of data samples with tags and without tags to realize semi-supervised learning; a loss function term considering error minimization of source domain data and target domain data to realize transfer learning; and norm regularizers constraining weight space. The domain transfer extreme learning machine method provide by the invention is combined with the source domain to process the problem of prediction of the target domain, thereby increasing the generalization capability and range of application of the extreme learning machine. Introduction of the manifold regularization term also enables the method proposed by the invention to still maintain a relatively good learning effect when data with tags are little, the restriction that a traditional machine learning method requires a large amount of data with tags is overcome, and the accuracy and robustness of prediction are also improved.

Description

technical field [0001] The invention relates to a field transfer extreme learning machine algorithm based on manifold regular items and Lp norm regularizers, and belongs to the technical fields of machine learning and pattern recognition. Background technique [0002] Artificial neural network has been widely used in various fields such as biology, chemistry, medicine, economy and ocean because of its powerful self-adaptation, self-organization, self-learning and nonlinear mapping capabilities. However, the traditional neural network, such as the Back Propagation (BP) network, needs to manually set a large number of network training parameters, the training speed is slow, and it is easy to generate a local optimal solution. In response to the above problems, Huang proposed a new algorithm for Single-hidden Layer Feedforward Neural Networks (SLFNs) called Extreme Learning Machine (ELM). The core of the algorithm is mainly two parts: one is to randomly generate input weights ...

AI Technical Summary

Problems solved by technology

However, traditional ELM requires a large amount of labeled data to train the classifier, and requires the training data and the target to be recognized to have the same distribution characteristics, that is, it does not have domain transfer capabilities, etc.

In real life, labeling the data will consume a lot of manpower and material resources, and the target to be identified and the training data may not necessarily meet the condition of the same distribution. Seawater turbidity, geological characteristics, etc.) and other factors, the images collected by AUV in two different sea areas on the seabed will have data offset phenomenon and the training sample data with prior knowledge (labeled data) is often a small amount

Images