Open domain-based transfer learning method and system

Abstract

The invention provides an open domain-based transfer learning method and system. The method comprises the steps of inputting a source domain sample set and a target domain sample set to a target neural network, thereby performing transfer learning on the target domain sample set, wherein a tag space corresponding to the target domain sample set is a sub-space of a tag space corresponding to the source domain sample set, and the target neural network is used for screening out a first sample set with tags belonging to a target domain and a source domain and a second sample set with tags only belonging to the source domain in the source domain sample set, facilitating positive transfer generated by the first sample set and inhibiting negative transfer generated by the second sample set. According to the open domain-based transfer learning method and system provided by the invention, the problem of transfer learning in an open domain is effectively solved.

Description

technical field [0001] The invention belongs to the technical field of computer data analysis, and more specifically, relates to a transfer learning method and system in an open field. Background technique [0002] Deep learning methods have achieved good results in machine learning problems and applications such as computer vision and natural language processing by training deep networks on large amounts of labeled data. Machine learning can solve problems in many different fields. However, large-scale labeled data sets are limited in number and application fields, and manually labeling a large amount of training data in various application fields requires a high cost. One of the solutions to the above problems An effective machine learning method is transfer learning. Transfer learning is a paradigm for learning discriminative models when there is a distribution shift between the source domain sample set and the target domain data. Transfer learning attempts to build a le...

AI Technical Summary

Problems solved by technology

[0004] In the case of such an open domain, the samples corresponding to the label classes (outliers) in the source domain that do not belong to the label space of the target domain will mislead the learner during domain matching, hindering the transfer of effective knowledge, making Existing transfer learning methods cannot achieve good results

Images