Entity and relationship joint extraction method based on reinforcement learning

Abstract

The invention discloses a joint information extraction method. According to the method, entity and relationship information is jointly extracted from an input text, and the method is composed of a joint extraction module and a reinforcement learning module. The joint extraction module adopts an end-to-end design and comprises a word embedding layer, a coding layer, an entity identification layer and a joint information extraction layer. Wherein the word embedding layer adopts a mode of combining a Glove pre-training word embedding library and word embedding representation based on character granularity. The encoding layer encodes the input text by using a bidirectional long-short memory network. The entity identification layer and the joint information extraction layer decode by using a one-way long-short memory network. The reinforcement learning module is used for removing noise in a data set, and a strategy network of the reinforcement learning module is composed of a convolutionalneural network. The strategy network comprises a pre-training process and a re-training process, and in the pre-training process, a pre-training data set is used for carrying out supervised training on the strategy network. In the retraining process, the strategy network is updated by obtaining the rewards of the joint extraction network, which is an unsupervised learning process.

Description

technical field [0001] The invention designs a method for jointly extracting entities and relationships based on reinforcement learning, which belongs to the technical field of natural language processing. Background technique [0002] In natural language processing tasks, entity recognition and relationship extraction are key technologies for building knowledge graphs, and are also important components of natural language processing applications such as semantic analysis, intelligent question answering, and automatic summarization. The core of this task is to extract entities of a given text and the semantic relationship between two entities. With the development of artificial intelligence and knowledge graph technology, this field has attracted the attention of more and more scholars. [0003] Traditional methods usually treat entity recognition and relation extraction tasks as two independent tasks, namely named entity recognition (Named Entity Recognition, NER) and rela...

AI Technical Summary

Problems solved by technology

[0006] Aiming at the problem of noisy data in the data set, the present invention proposes a new network model for joint entity and relation extraction

Images