Automatic Detection of Research Values from Scientific Abstracts Across Computer Science Subfields

2025-02-23

AI Summary

This paper addresses the automatic identification of research values across multiple subfields of computer science (CS). We propose the first comprehensive, ten-dimensional value annotation scheme covering the entire CS domain and construct a large-scale detection framework spanning 32 subfields and 86 conferences/journals. Methodologically, leveraging fine-grained human-annotated data, we integrate domain-adaptive text representations with a multi-label classification model to automatically identify values in 226,000 abstracts from the past decade. Our key contributions are: (1) the first systematic definition and large-scale annotation of domain-general research values in CS; (2) empirical characterization of cross-subfield value distribution disparities and longitudinal evolution over ten years; and (3) an extensible, value-oriented analytical paradigm that provides empirical foundations for research policy formulation and disciplinary development assessment.

Technology Category

• Application Domains: Humanities & Computational Social Science
• Humans and AI: Learning Human Values and Preferences
• Cognitive Modeling & Cognitive Systems: Affective Computing

Application Category

• Economics, Online Markets and Human Computation: Research challenges in human and human-AI computation
• Web Mining and Content Analysis: Interdisciplinary science discovery with web data mining
• Social Networks and Social Media: Computational social science

Abstract

The field of Computer science (CS) has rapidly evolved over the past few decades, providing computational tools and methodologies to various fields and forming new interdisciplinary communities. This growth in CS has significantly impacted institutional practices and relevant research communities. Therefore, it is crucial to explore what specific extbf{research values}, known as extbf{basic and fundamental beliefs that guide or motivate research attitudes or actions}, CS-related research communities promote. Prior research has manually analyzed research values from a small sample of machine learning papers cite{facct}. No prior work has studied the automatic detection of research values in CS from large-scale scientific texts across different research subfields. This paper introduces a detailed annotation scheme featuring extbf{ten research values} that guide CS-related research. Based on the scheme, we build value classifiers to scale up the analysis and present a systematic study over 226,600 paper abstracts from 32 CS-related subfields and 86 popular publishing venues over ten years.

Problem

• Automatic detection of research values
• Analysis across computer science subfields
• Large-scale scientific text classification

Innovation

• Automatic detection of research values
• Large-scale scientific text analysis
• Value classifiers for CS subfields