Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Extracting Abstraction Dimensions by Identifying Syntax Pattern from Texts

Jian Zhou, Jiazheng Li, Sirui Zhuge, Hai Zhuge

TL;DR

This work introduces a syntax-pattern–driven method to automatically discover four abstraction dimensions—subject, action, object, and adverbial—from texts, structuring them into four abstraction trees that compose a multi-dimensional resource space for text querying. Subclass relations among patterns are inferred via six formal rules, enabling complete and independent dimension trees (1NF/2NF/3NF) that cover the majority of textual content. The approach supports natural-language question answering by mapping questions to the corresponding four dimensions and retrieving candidate sentences from multiple trees, outperforming traditional ranking-based and some LLM-based methods in terms of coverage and relevance. The method is validated across six diverse datasets, achieving strong precision/recall/F1 on subclass extraction and demonstrating high QA precision on real queries, illustrating its interpretability, scalability, and applicability to diverse domains without external ontologies.

Abstract

This paper proposed an approach to automatically discovering subject dimension, action dimension, object dimension and adverbial dimension from texts to efficiently operate texts and support query in natural language. The high quality of trees guarantees that all subjects, actions, objects and adverbials and their subclass relations within texts can be represented. The independency of trees ensures that there is no redundant representation between trees. The expressiveness of trees ensures that the majority of sentences can be accessed from each tree and the rest of sentences can be accessed from at least one tree so that the tree-based search mechanism can support querying in natural language. Experiments show that the average precision, recall and F1-score of the abstraction trees constructed by the subclass relations of subject, action, object and adverbial are all greater than 80%. The application of the proposed approach to supporting query in natural language demonstrates that different types of question patterns for querying subject or object have high coverage of texts, and searching multiple trees on subject, action, object and adverbial according to the question pattern can quickly reduce search space to locate target sentences, which can support precise operation on texts.

Extracting Abstraction Dimensions by Identifying Syntax Pattern from Texts

TL;DR

This work introduces a syntax-pattern–driven method to automatically discover four abstraction dimensions—subject, action, object, and adverbial—from texts, structuring them into four abstraction trees that compose a multi-dimensional resource space for text querying. Subclass relations among patterns are inferred via six formal rules, enabling complete and independent dimension trees (1NF/2NF/3NF) that cover the majority of textual content. The approach supports natural-language question answering by mapping questions to the corresponding four dimensions and retrieving candidate sentences from multiple trees, outperforming traditional ranking-based and some LLM-based methods in terms of coverage and relevance. The method is validated across six diverse datasets, achieving strong precision/recall/F1 on subclass extraction and demonstrating high QA precision on real queries, illustrating its interpretability, scalability, and applicability to diverse domains without external ontologies.

Abstract

This paper proposed an approach to automatically discovering subject dimension, action dimension, object dimension and adverbial dimension from texts to efficiently operate texts and support query in natural language. The high quality of trees guarantees that all subjects, actions, objects and adverbials and their subclass relations within texts can be represented. The independency of trees ensures that there is no redundant representation between trees. The expressiveness of trees ensures that the majority of sentences can be accessed from each tree and the rest of sentences can be accessed from at least one tree so that the tree-based search mechanism can support querying in natural language. Experiments show that the average precision, recall and F1-score of the abstraction trees constructed by the subclass relations of subject, action, object and adverbial are all greater than 80%. The application of the proposed approach to supporting query in natural language demonstrates that different types of question patterns for querying subject or object have high coverage of texts, and searching multiple trees on subject, action, object and adverbial according to the question pattern can quickly reduce search space to locate target sentences, which can support precise operation on texts.
Paper Structure (29 sections, 3 figures, 9 tables)

This paper contains 29 sections, 3 figures, 9 tables.

Table of Contents

  1. Introduction
  2. Motivation Example
  3. General Idea for Extracting Subclass Relations
  4. General Process for Extracting Dimension and Query
  5. Extracting Subclass Relations
  6. Identifying Patterns of Subject, action, object, and adverbial
  7. Identifying Subclass Relations
  8. Completeness
  9. Patterns of Question and Patterns of Answer
  10. Patterns of Question
  11. Properties of Answer
  12. Patterns of Answer
  13. Completeness of the Patterns of Question and Answer
  14. Relations between Questions and Answers
  15. Verification of Extracted Class Trees
  16. ...and 14 more sections

Figures (3)

  • Figure 1: Four syntax dimensions extracted from the short input texts.
  • Figure 2: Examples of extracting dimensions according to syntax and query resource space.
  • Figure 3: General process of extracting four dimensions from Summ dataset and process of querying the space with the four dimensions in natural language.