Apparatus and method for embedding sentence feature vector

Abstract

An apparatus and method for embedding a sentence feature vector are disclosed. An apparatus for embedding a sentence feature vector according to an embodiment includes a sentence acquisitor configured to acquire a first sentence and a second sentence, each including more than one word; a vector extractor configured to extract a first feature vector corresponding to the first sentence and a second feature vector corresponding to the second sentence by independently inputting each of the first sentence and the second sentence into a feature extraction network; and a vector compressor configured to compress the first feature vector and the second feature vector into a first compressed vector and a second compressed vector, respectively, by independently inputting each of the first feature vector and the second feature vector into a convolutional neural network (CNN)-based vector compression network.

Description

technical field [0001] The disclosed embodiments relate to a technique of embedding feature vectors representing features of sentences. Background technique [0002] Recently, due to the advancement of artificial intelligence (AI: Artificial Intelligence) in the field of natural language processing, a number of natural language models that perform embedding at the word level have been disclosed. [0003] However, in order to provide meaningful services using natural language processing, it is necessary to efficiently perform operations such as machine reading comprehension (Machine Reading Comprehension) or sentence similarity (Semantic Textual Similarity) comparison, while only embedding at the word level to handle such Operation is limited. [0004] In order to overcome such problems, several network structures capable of performing sentence-level embeddings have been proposed, however, there are still service providers related to natural language processing that are diff...

AI Technical Summary

Problems solved by technology

[0003] However, in order to provide meaningful services using natural language processing, it is necessary to efficiently perform operations such as machine reading comprehension (Machine Reading Comprehension) or sentence similarity (Semantic Textual Similarity) comparison, while only embedding at the word level to handle such operation is limited

[0004] In order to overcome such problems, several network structures capable of performing sentence-level embeddings have been proposed, however, there are still service providers related to natural language processing that are difficult to utilize alone due to the large amount of time and resources required for learning. The problem

Images