Natural Language Processing Using Semantic Compatible Relations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional natural language processing tools face difficulties in handling languages with loose grammars, such as Chinese, due to frequent omissions, changes in sentence order, and ungrammatical sentences, leading to errors in semantic analysis and machine translation.
Innovation Solution
A method for natural language processing that reduces input words into semantic structures using dependence parse trees based on compatible relations stored in a database, considering semantic context and ontology to improve parsing accuracy and reduce computational complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional grammar-based parsing systems are used, then they work well for languages with strict grammar rules, but they fail to handle languages with loose grammars like Chinese that have omissions, changes of order and ungrammatical sentences
Solution Approach 1:
The patent changes the fundamental parameters of parsing from grammar-based rules to semantic-based compatible relations. Instead of relying on grammatical structures that vary across languages, the system uses semantic compatibility relationships that are language-agnostic, allowing the same parsing approach to work reliably across languages with different grammatical characteristics.
Solution Approach 2:
The patent substitutes the mechanical grammar-based parsing system with a semantic relation-based system. Rather than applying rigid grammatical rules mechanically, the system uses semantic compatibility relationships stored in the database to determine parsing structures, replacing the mechanical grammar-checking approach with a more flexible semantic matching approach.
2Reliability
If semantic information is considered to correctly parse ungrammatical sentences, then parsing accuracy improves, but computational complexity increases due to needing to consider all possibilities
Solution Approach 1:
The patent performs preliminary action by pre-computing and storing semantic compatible relations in a database before the actual parsing task. This pre-processing step creates a lookup table of valid semantic relationships, so during parsing, the system only needs to query this pre-computed data rather than calculating all possible semantic combinations in real-time, significantly reducing computational complexity.
Solution Approach 2:
The patent segments the parsing problem into two independent parts: (1) word segmentation using compatible relations from the database, and (2) parse tree construction using these segmented words. This segmentation allows each sub-problem to be solved independently and efficiently, avoiding the need to consider all possibilities in a single complex computation.
3Measurement precision
If all possible semantic combinations are considered to ensure accurate parsing, then parsing precision improves, but execution time increases due to exhaustive search
Solution Approach 1:
The patent pre-computes semantic compatible relations and stores them in the database before execution. During actual parsing, the system queries this pre-computed data rather than calculating all semantic combinations in real-time, achieving both high precision and fast execution by separating the computationally intensive pre-processing from the time-sensitive parsing operation.
Solution Approach 2:
The patent creates a copy of semantic relationship data in the form of a database containing pre-computed compatible relations. Instead of repeatedly calculating semantic compatibility during parsing, the system copies the necessary relationship data into the database structure, allowing rapid retrieval and comparison during the parsing process without re-computing semantic relationships.
4Device complexity
If recursive reduction method is used, then computational complexity decreases compared to exhaustive search, but the method requires deterministic execution results which may limit handling of ambiguous cases
Solution Approach 1:
The patent incorporates feedback by using the database of compatible relations to validate and guide the recursive reduction process. At each step of the recursive parsing, the system queries the database to receive feedback on whether the current parsing choice is semantically valid, allowing the deterministic algorithm to correct itself and handle ambiguous cases by referring back to the pre-computed semantic relationships.
Data Source
AI summary
A natural language processing method, comprising: receiving multiple input words; and reducing the multiple words into one or more subject word data structures according to sets stored in a database, wherein one of the subject word data structures includes a first input word and a second input words among the input words, wherein one of the sets includes a compatible relation between the first input word and the second input word, wherein the compatible relation between the first input word and the second word includes a compatible property for denoting an intensity representing occurrences of the first input word and the second word in a training corpus.