Model fusion triad representation learning system and method based on deep learning

Abstract

The invention discloses a model fusion triad representation learning system and method based on deep learning. The method comprises the following steps: carrying out the embedded representation of a word through a pre-trained BERT language model, and obtaining a more contextualized representation of the word; meanwhile, a masking language modeling task of a BERT structure is used for taking a triple of the masking language modeling task as sequence input; the method is used for solving the problem of multiple semantics of the same entity; the mapping entity relationship can be represented differently in different fields by using a projection or conversion matrix; however, the transformed BERT can take the triad or the description information thereof as text input and train the triad and the description information together; the mechanism of the BERT itself has different word vectors for the entity relationship in different sentences, and the problem of different semantics of the entityrelationship is effectively solved, so that the selection of TransE is not limited by the model itself. On the contrary, the model is simple enough to truly reflect the corresponding relationship among the triples. Meanwhile, the complexity of the model is reduced.

Description

【Technical field】 [0001] The invention belongs to the field of electronic information technology, and relates to a model fusion triple representation learning system and method based on deep learning. 【Background technique】 [0002] People usually organize the knowledge in the knowledge base in the form of a network, each node in the network represents an entity (person name, place name, organization name, concept, etc.), and each edge represents the relationship between entities. Therefore, most knowledge can often be represented by triples (entity 1, relation, entity 2), corresponding to an edge in the knowledge base network and the two connected entities. This is a common representation of knowledge bases. For example, the resource description framework (RDF) technical standard released by the World Wide Web (W3C) is based on triple representation. Knowledge base is an important basic technology to promote the development of artificial intelligence disciplines and suppor...

AI Technical Summary

Problems solved by technology

It is also possible to learn the representation and learning of entities and relationships in the knowledge graph by using the information of the relationship path. The disadvantage and deficiency is that the selection and design of the path formula directly determines the performance of the model, which is insufficient in versatility.

Images