Semantic Similarity Measure of Natural Language Text through Machine Learning and a Keyword-Aware Cross-Encoder-Ranking Summarizer - A Case Study Using UCGIS

Aug 6, 2026

Upload by Jiven

This study focuses on improving topic relationship identification within GIS&T Body of Knowledge (BoK), a community-driven geospatial knowledge effort initiated by UCGIS. Because current topic links are often created manually, semantic relationships such as semantic similarity may be incomplete. The work evaluates multiple NLP approaches for extracting meaning from text, comparing deep neural networks and traditional machine learning. It also proposes KACERS, a keyword-aware cross-encoder ranking summarizer, to produce semantic summaries of scientific publications and support automated topic navigation, with emphasis on long-text semantic understanding.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages23

Reading Time58 min

AccessFree to download

Abstract

Introduction

GIS&T BoK background and motivation

Semantic similarity and NLP techniques

Need for automation and handling long texts

Semantic Similarity Measure of Natural Language Text through Machine Learning and a Keyword-Aware Cross-Encoder-Ranking Summarizer-A Case Study Using UCGIS GIS&T Body of Knowledge Yuanyuan Tian 1, Wenwen Li 1,*, Sizhe Wang 1, Zhining Gu 1 1School of Geographical Sciences and Urban Planning, Arizona State University, Arizona, Tempe, USA *[Corresponding Author: ](Corresponding Author: wenwen@asu.edu)[wenwen@asu.edu](Corresponding Author: wenwen@asu.edu) Abstract: Initiated by the University Consortium of Geographic Information Science (UCGIS), GIS&TBody of Knowledge (BoK) is a community-driven endeavor to define, develop, and document geospatial topics related to geographic information science and technologies (GIS&T) . In recent years, GIS&T BoKhas undergone rigorous development in terms of its topic re-organization and content updating, resulting in a new digital version of the project. While the BoK topics provide useful materials for researchers and students to learn about GIS, the semantic relationships among the topics, such as semantic similarity, should also be identified so that a better and automated topic navigation can be achieved. Currently, the related topics are either defined manually by editors or authors, which may result in an incomplete assessment of topic relationship. To address this challenge, our research evaluates the effectiveness of multiple natural language processing (NLP) techniques in extracting semantics from text, including both deep neural networks and traditional machine learning approaches. Besides, a novel text summarizationKACERS (Keyword-Aware Cross-Encoder-Ranking Summarizer) -is proposed to generate a semantic summary of scientific publications. By identifying the semantic linkages among key topics, this work provides guidance for future development and content organization of the GIS&T BoK project. It also offers a new perspective on the use of machine learning techniques for analyzing scientific publications, and demonstrate the potential ofKACERS summarizer in semantic understanding of long text documents. 1. INTRODUCTION A Body of Knowledge (BoK) defines important knowledge topics to complete a job or task in a specific domain, and contributes to professional development needs (Stelmaszczuk-Gorska et al., 2020) . Expected users of a BoK include educators, students, and professionals. GIS related BoK works can assist GIS domain curriculum planning and revision, program assessment and articulation, professional certification, and employee screening (Prager & Plewe, 2009; C. Wang et al., 2020). For example, GIS&T Body of Knowledge, or GIS&T BoK for short, represents the Geographic Information Science and Technology domains in a hierarchical fashion of knowledge areas, units, and topics. GIS&T BoK project is adopted by the American Association of Geographers as a standard GIScience learning guideline (DiBiase et al., 2007) . Encyclopedia-style BoKs have begun to include explicit cross-references or a list of related topics at the end of each topic, which makes it easier for knowledge seekers to explore related learning materials. This practice is common, just like in Wikipedia’s “see also” section, the “cross-reference”section also exists in the Encyclopedia of GIS (Shekhar & Xiong, 2007) , Encyclopedia of Geographic Information Science (K. Kemp, 2008), International Encyclopedia of Geography (Brunn, 2019) , and GIS&T BoK. Topic relevance can be manually crafted, but it is possible to automate the process by Natural Language Processing (NLP). NLP is a technique to process and analyze language data, aiming to extract meanings (semantics) from text (Nadkarni et al., 2011) . In recent years, the advances in Geospatial Artificial Intelligence (GeoAI) (Li, 2020, 2022) have further empowered NLP to utilize advanced AI technology especially deep learning to process and understand natural language text. For instance, one of the main research applications of NLP is to measure the s

FAQs

Why is automated semantic relationship identification needed for GIS&T BoK?

Current topic cross-references are often defined manually, which can lead to incomplete assessments of topic relationship and require substantial effort to analyze long publications.

How does the research evaluate NLP methods for extracting semantics?

It evaluates multiple NLP techniques, covering both deep neural network approaches and traditional machine learning methods, to extract semantic content from text.

What is KACERS and what is its purpose?

KACERS is a keyword-aware cross-encoder ranking summarizer proposed to generate semantic summaries of scientific publications by leveraging semantic linkages among key topics.

Aug 6, 2026

Upload by Jiven