Coding method, device and equipment, and storage medium

Abstract

The embodiment of the invention discloses coding method, device and equipment, and a storage medium, and belongs to the field of natural language processing. The coding method comprises the steps: obtaining a target paragraph and a context statement of the target paragraph, and inputting the target paragraph and the context statement into a memory coding model; and obtaining an original vector setand a memory vector set in an input layer, obtaining a first target statement matrix of the original vector set in a memory layer according to the original vector set and the memory vector set, obtaining a paragraph vector of a target paragraph according to the first target statement matrix in an output layer, and carrying out processing based on the paragraph vector. And each statement is codedby taking the target paragraph as a unit, so that the coding speed is increased. Moreover, in the coding process, the target paragraph itself is considered; the context statement of the target paragraph is also considered; the paragraph vector for accurately describing the target paragraph can be obtained according to the association relationship between the statement in the target paragraph and the context statement; and the coding accuracy is improved.

Description

technical field

[0001] The present invention relates to the field of natural language processing, and in particular, to an encoding method, apparatus, device and storage medium. Background technique

[0002] Encoding is the process of converting text into encoded values, resulting in a vector that accurately describes the meaning of the text. By encoding, the text can be converted into a vector form that is convenient for operation and processing, and it has been widely used in many fields such as sentence selection and sentence generation.

[0003] At present, a hierarchical encoding scheme is proposed. When a target paragraph including multiple sentences is to be encoded, the word vector of each word in each sentence in the target paragraph is obtained. For each sentence in the target paragraph, the first encoding model based on the word level is applied, and the word vector of each word in the sentence is encoded into a vector to obtain the sentence vector of the sentenc...

Images