Natural Language Processing Engine for Nested Clause Analysis
Solution Overview
Problem
Existing natural language processing methods, such as Latent Semantic Indexing and SAO/eSAO triples, fail to effectively capture the complexity of natural language documents, particularly neglecting nested clauses and syntactic nuances, leading to incomplete information retrieval and search results.
Innovation Solution
A natural language processing engine (NLPE) that analyzes input sentences to generate multiple syntactic structures at sentence, phrase, and entity levels, recognizing main and subordinate clauses, and linking entities to an external knowledge base to create semantic structures and tuples for improved search and information retrieval.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If Latent Semantic Indexing is used to extract keywords and represent documents, then information retrieval can be performed efficiently, but meaningful features and syntactic information of the document are discarded or failed to be considered
Solution Approach 1:
The patent segments the document analysis process into multiple levels: sentence-level parsing to identify clauses, phrase-level analysis to extract syntactic structures, and entity-level identification to recognize meaningful features. This segmentation allows the system to retain both efficiency and meaningful information by processing documents hierarchically rather than as a whole.
Solution Approach 2:
The patent adds syntactic and semantic dimensions to the traditional keyword-based approach. By constructing SAO tuples with subject, action, and object components, and further enriching them with entity types and relationships, the system transforms flat keyword representations into multi-dimensional semantic structures that preserve meaningful features while maintaining retrieval efficiency.
2Loss of information
If Natural Language Processing is used to extract SAO triples from sentences, then more syntactic information can be obtained, but the complexity of linguistics such as nested clauses within a sentence is not addressed
Solution Approach 1:
The patent applies segmentation by breaking down complex sentences into constituent clauses and phrases. The sentence parsing module divides sentences into main clauses and subordinate clauses, the phrase parsing module further segments phrases into syntactic components, and the entity-identification module isolates specific entities. This hierarchical segmentation makes it feasible to handle linguistic complexity by processing each segment separately rather than attempting to analyze the entire sentence structure at once.
3Measurement precision
If multiple parsing modules are executed to generate syntactic structures and tuples, then more accurate semantic understanding is achieved, but processing time and computational complexity increase
Solution Approach 1:
The patent implements preliminary action by performing sentence-level parsing first to identify the overall clause structure, then using this information to guide phrase-level analysis. The syntactic structures generated at each level are reused in subsequent processing stages, avoiding redundant analysis. This preliminary structuring reduces the computational burden of later entity identification and tuple generation, thereby reducing overall processing time while maintaining accuracy.
Data Source
AI summary
The subject disclosure presents a natural language processing engine for analyzing an input sentence comprising one or more clauses, and generating a plurality of semantic structures for the sentence and the component clauses, based on statistically parsing an input sentence to generate a syntactic structure of the input sentence, examining the syntactic structure of phrases and subordinate clauses within the input sentence, and generating tuples representing a subject, verb, object, indirect object, supplement, type, etc. Each part of the tuple is a reference to an entity in an external knowledge base. Disclosed operations include linking a plurality of entities identified in the syntactic structure with corresponding entities found in an external knowledge base, operating a co-reference resolution, filtering the references from mentioned entities to external entities by semantic relations, and exporting the set of output tuples.