Natural Language Processing Distributed Vector Calculation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing natural language processing techniques fail to comprehensively represent the purpose of a sentence associated with a classification word, even when using distributed vectors for words or phrases as tags, as they only represent aspects of the background or characteristics of classified sentences.
Innovation Solution
A natural language processing apparatus that calculates distributed vectors for classification words by leveraging distributed vectors of words and sentences, allowing for comprehensive representation of the purpose of sentences through a three-unit system: one for word vectors, one for sentence vectors, and one for classification word vectors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If distributed vectors of words or phrases are used as distributed vectors of tags, then semantic comparison of natural language can be performed, but the purpose of the sentence associated with the tag is not comprehensively represented
Solution Approach 1:
The patent merges three types of distributed vectors (word vectors, sentence vectors, and tag name vectors) through weighted summation to create a comprehensive tag distributed vector. This combining approach integrates multiple information sources to fully represent both the semantic meaning and the sentence purpose associated with each tag, thereby resolving the information loss problem while maintaining semantic comparison capabilities
2Device complexity
If only tag name distributed vectors are used, then computational complexity is reduced, but the representation of sentence purpose becomes incomplete
Solution Approach 1:
The patent implements a configurable weighted summation approach where the contribution of each vector type (word, sentence, tag name) can be adjusted through weights. This allows the system to perform partial computations when full comprehensiveness is not required, or to use all three components when complete sentence purpose representation is needed, thus providing flexibility to balance computational complexity against information completeness
Data Source
AI summary
A natural language processing apparatus includes: a first calculation unit configured to calculate a distributed vector of a word included in a plurality of sentences based on a database that manages the plurality of sentences associated with a classification word; a second calculation unit configured to calculate a distributed vector of the sentence based on the distributed vector of the word included in each sentence; and a third calculation unit configured to calculate a distributed vector of the classification word based on the distributed vector of each sentence associated with the same classification word.