A knowledge map reasoning algorithm based on a stacked neural network

Abstract

The invention discloses a knowledge map relation inference algorithm based on a stacked neural network, and belongs to the technical field of artificial intelligence representation learning. The invention constructs a stacked neural network model comprising two components: a standard LSTM loop neural network and a multi-layer perceptron network. The model regards the triplets in the knowledge mapas short sentences and uses the learning ability of LSTM loop neural network to model the logical and semantic characteristics of the knowledge map in order to learn the grammatical and semantic information in the knowledge map. Through the bottom-level feature learning process, the upper-level fully connected network can provide enough discrimination to distinguish objects in different contexts.The algorithm of the invention models the three tuples in the knowledge map from the perspective of semantics, fundamentally solves the semantic diversity of entities and relationships, can greatly reduce the computational overhead of the relational reasoning of the large-scale knowledge map, and simultaneously can ensure the effectiveness of the relational reasoning.

Description

technical field [0001] The invention belongs to the technical field of artificial intelligence representation learning, and in particular relates to a knowledge map relation reasoning algorithm based on a stacked neural network. Background technique [0002] Logical reasoning, also known as multi-relational reasoning in the field of statistical relational learning, has been a major problem in artificial intelligence research. It is well known that the ability of knowledge to infer new facts is of great importance in many artificial intelligence applications, especially in knowledge-based expert systems. Recent advances in knowledge graphs have reshaped our understanding of logical reasoning, from logical rule-based approaches to a broader statistical relational learning perspective. Among them, the translation model represented by TransE has so far been the most cited multi-relational data modeling work, attracting the interest of a large number of researchers. [0003] Ho...

AI Technical Summary

Problems solved by technology

However, the linear change does not change the basic idea of ​​the TransE model

The literature "Transition-based Knowledge Graph Embedding with Relational Mapping Properties" introduces weight factors related to relation types into the optimization objective function to strengthen the TransE model's ability to distinguish entities, but it still cannot fundamentally solve the problem of TransE being affected by semantic diversity. the impact of

Images