白丝美女被狂躁免费视频网站,500av导航大全精品,yw.193.cnc爆乳尤物未满,97se亚洲综合色区,аⅴ天堂中文在线网官网

Generation of sentence representation

專利號
US11176333B2
公開日期
2021-11-16
申請人
International Business Machines Corporation(US NY Armonk)
發(fā)明人
Bang An; HongLei Guo; Shiwan Zhao; Zhong Su
IPC分類
G06F40/56; G06F40/30; G06F40/289; G06F40/58
技術(shù)領(lǐng)域
sentence,graph,word,nodes,node,syntactic,in,neighbor,may,cloud
地域: NY NY Armonk

摘要

Embodiments of the present disclosure relate to generation of sentence representation. In an embodiment, a method is disclosed. According to the method, a sentence graph is generated from a sentence containing words, the sentence graph comprising nodes representing the words and edges connecting the nodes to indicate relationships between the words. Word representations for the plurality of words are determined based on the sentence graph by applying a graph convolution operation on respective sets of neighbor nodes for respective ones of the nodes, a set of neighbor nodes for a node having edges connected with the node. A sentence representation for the sentence is determined based on the word representations for use in a natural language processing task related to the sentence. In other embodiments, a system and a computer program product are disclosed.

說明書

The word representation for each of the words in the sentence 402 may be determined by the word representation generation module 520. All the word representations 528 may be provided for the representation generation module 530. The representation generation module 530 determines, based on the word representations 528, a sentence representation 412 for the sentence 402 for use in a natural language processing task related to the sentence. In some embodiments, the word representations 528 may be organized together to directly form the sentence representation 412. In some other embodiments, the representation generation module 530 may further process the word representations, for example, by applying one or more other neural network layers. The scope of the embodiments of the present invention is not limited in this regard.

The sentence representation 412 may be utilized in various manners in different natural language processing tasks, such as by the decoder 420 of the system 400. The utilization of the sentence representation 412 is also not limited in the embodiments of the present invention. Some examples of the natural language processing tasks may include machine translation, NLI, semantic role labeling, entity reorganization, text summarization, reading comprehension, relation extraction, and so on.

權(quán)利要求

1
微信群二維碼
意見反饋