Call for Book Chapters

How to Use LLMs in Your Research

A Practical Guide for Science of Science,
as well as STI Policy, Innovation Management and Economics

Book Description

This edited book invites contributions from leading experts who have applied Large Language Models (LLMs) in their research within the fields of science of science, science, technology and innovation (STI) policy, innovation management and economics. Prioritising methodological clarity, it aims to equip researchers with detailed knowledge on leveraging LLMs-based techniques. It adopts a ‘how-to’ and knowledge transfer approach, focusing on explaining the LLM methods used to address research questions or challenges, and their rationale.

Chapters will provide detailed explanations of LLM methods, clarifying the types of research questions they can and cannot address, and discussing their potential and limitations. Where possible, chapters will include insets with illustrative examples or case studies, facilitating reader comprehension. The book will be accompanied by a GitHub repository containing articles, documents, code, and data for selected studies, enabling hands-on learning and replication. We also seek introductory chapters covering the foundational and theoretical aspects of LLMs.

By bridging theory and practice, ‘How to Use LLMs in Your Research’ aims to empower researchers to harness the potential of LLMs responsibly and effectively. It contributes to standardising procedures and establishing ethical guidelines, underscoring human oversight and accountability in AI-integrated research. Ultimately, it aims to equip researchers with the knowledge and skills for impactful and reliable research. To this end, chapters must be written in a clear and accessible style.

Rationale for the book

The rapid development of AI, especially in Natural Language Processing (NLP), is transforming research methods across disciplines, particularly in Social Sciences (SS), where it enables analysis of massive unstructured textual data. Despite this potential, SS researchers often hesitate to embrace AI tools due to steep learning curves and lack of clear and established guidelines. However, the recent surge in user-friendly LLMs platform, like ChatGPT, has significantly altered the landscape, presenting both unprecedented opportunities and unique challenges.

The democratisation of AI access while beneficial, introduces critical challenges for researchers. Without a sufficient grasp of LLM operational principles and inherent limitations, there is a risk of methodological misuse, which may compromise the validity and reliability of research findings. Thus, ‘How to Use LLMs in Your Research’ aims to provide guidance for researchers integrating LLMs-based techniques. It focuses on methodological illustration and applications within the broad fields of science of science, and science, technology and innovation (STI) policy, innovation management and economics.

Target Audience

‘How to Use LLMs in Your Research’ is tailored for researchers seeking to enhance their methodologies by integrating LLMs, providing methodological guidance and real-world examples. Academics aiming to remain at the forefront of AI-driven research will find this book essential for understanding the transformative potential of LLMs in their respective fields. Policy-makers, who increasingly interact with chief data scientists and their teams, will greatly benefit from this book. They will learn the common language of computational semantics and gain awareness of what can, and cannot, be achieved by LLMs in research and its applications.

Invitation to Contribute /Submissions Invited

This book welcomes two types of contributions:

1. Methodological Chapters (~70% of the book): These chapters should provide detailed explanations of specific LLM-based methodologies, offering an easy-to-understand resource to be used as a guide by readers for their research. These chapters should clearly highlight strengths and weaknesses when applied to various research questions and data types. We strongly encourage authors to include an appendix (500-1,000 words) to provide a concise yet illustrative example of how the methodology has been employed in a real-world research scenario, and/or policy-related design or application. This appendix could mention a case study, an empirical analysis, or a detailed walkthrough of a specific research task. These appendices will be used as illustrative inset boxes within the chapters. These methods must be applicable to research in the science of science, STI policy, innovation management and economics.
2. Theoretical and Critical Discussions (~30% of the book): To complement the methodological focus of the book, the editors also encourage submissions of chapters that will delve into the theoretical underpinnings of LLMs, examining topics such as the theory of LLM architectures, or the history of text mining. Contributions that critically analyse the current and future roles of LLMs in scientific research are also encouraged. This includes discussions on methodological implications, ethical considerations, potential societal impacts, and the evolving relationship between LLMs and human researchers.

Potential Topics for Chapters

The following topics provide a starting point for potential chapter contributions, with a particular emphasis on foundational understanding and methodological applications of LLMs within the social sciences. This list is not exhaustive, and we welcome innovative proposals that explore new avenues and perspectives.

I. Core Concepts and Foundations

• Understanding LLMs for social sciences
  • Defining LLMs and their underlying mechanisms (e.g. transformers, attention mechanisms)
  • Exploring the capabilities and limitations of LLMs in research (e.g. what they can do, be used for, and what they cannot do or be used for)
  • Comparing LLMs to traditional NLP techniques and their respective strengths and weaknesses
  • The evolution of LLMs and their impact on various research domains
• Designing research and reporting research results using LLMs
  • Hypothesis generation and testing with LLMs
  • Data collection, analysis, and visualisation with LLMs
  • Qualitative and quantitative research with LLMs
  • Reporting and disseminating research findings with LLMs

II. LLM methodologies

• Core LLM Methodologies
  • Prompt Engineering: This section should delve deeper into various prompting techniques, including:
    • Zero-shot, Few-shot, and Chain-of-Thought prompting
    • Prompt design for using pre-trained LLMs, with best practices and strategies
  • Fine-tuning and Adaptation: Be specific about different fine-tuning approaches:
    • Fine-tuning for specific domains or tasks
    • Parameter-efficient fine-tuning techniques (e.g., adapters, prompt tuning)
  • Multimodal Methods: Explore LLMs that integrate different modalities (e.g., text and images).
• LLM Applications in Research
  • Information Retrieval and Knowledge Extraction:
    • LLMs for entity extraction and summarisation
    • LLMs for performing literature reviews and topic extraction
    • Embeddings and Similarity Search for clustering, classification, and information retrieval
  • Text Analysis and Understanding:
    • Named Entity Recognition (NER) and Relation Extraction
    • Sentiment Analysis and Opinion Mining
  • Hypothesis Generation and Causal Inference:
    • Generative LLMs for Hypothesis Generation
    • LLMs for Causal Inference
  • Advanced LLM Techniques
    • Explainable LLMs: Techniques for interpreting and explaining LLM outputs, addressing the « black box » problem.

III. Future Directions and Discussion

• Ethical considerations and broader discussion
  • Addressing biases, fairness, and transparency in LLMs and their outputs
  • Ensuring responsible and ethical development and deployment of LLMs in research
  • Establishing ethical guidelines and best practices for using LLMs
  • The role of human oversight and accountability in LLM-driven research
• Emerging methodological trends
  • Exploring the latest advancements and future potential of LLMs in research
  • Discussing the evolving role of LLMs in shaping research practices and methodologies
  • Proposing future research questions and directions in the field of LLMs and research
• Research directions
  • Identifying new areas of application for LLMs in science of science, STI policy, innovation management and economics
  • Encouraging interdisciplinary collaborations and knowledge sharing
  • Fostering the development of new LLM-based tools and techniques for research

Submission Guidelines

1. Long Abstract Submission

Potential contributors are invited to submit a long abstract (500 words, excluding the list of references, tables, and figures) that provides a comprehensive overview of the proposed chapter. The abstract should begin by summarising the chapter’s core content, with particular emphasis on the specific LLM-based method discussed and the types of research questions it can be applied to. It should clearly articulate how the method addresses these research questions.

If applicable, authors should indicate one or two possible examples of application of the methods that help to illustrate the practical relevance and potential impact of the proposed chapter.

Authors should also include a tentative title and 5-10 keywords that best represent the chapter’s theme.

The deadline for abstract submission is May 15, 2025.

2. Full Chapter Submission (upon invitation)

Upon review and acceptance of the long abstract, authors will be invited to submit a full chapter. Full chapters should be specifically prepared for this book. They should not exceed 7,000 words (including all references, appendices, tables, figures, etc.). Detailed formatting guidelines will be provided to authors upon acceptance of their abstracts.

3. Formatting Guidelines for Long Abstract:

• Use 1.5-line spacing and 11pt. Arial font.
• Include author affiliations and contact information on a separate page, named ‘Author Page.’
• Submit the abstract and the author page as two separate Word documents (.docx). The name of the two documents should have ‘_Abstract’ or ‘_Author_page’ as a suffix and the title of the work as the main part of the filename. For example: ‘LLM as tool_Abstract.docx’, and ‘LLM as tool_Author_page.docx’.
• Please submit to 4point0@polymtl.ca, using ‘[How to LLM book]:’ as a prefix in the subject line.

4. Review Process

All submitted abstracts and full chapters will undergo a double-blind peer-review process to ensure quality and relevance to the book’s theme. Authors will receive feedback from reviewers and may be asked to revise their chapters accordingly. Selected authors will be solicited to review the chapters of other contributors in order to ensure coherence and avoid excessive redundancy between chapters.

5. Important Dates

• 500-word abstract submission deadline: May 15, 2025
• Selection of abstracts and notification of acceptance: June 15, 2025
• Full chapter submission deadline: November 15, 2025
• Feedback from the reviewers: January 15, 2026
• Revised chapter submission deadline: May 15, 2026
• Feedback from the reviewers (second peer-review if necessary): June 15, 2026
• Manuscript delivery: July 15, 2026
• Approval of final proofs: July 31, 2026
• Publication (tentative date): August-Septembre 2026

6. Submission

Please submit your long abstract to 4point0@polymtl.ca. Use ‘[How to LLM book]:’ as a prefix in the subject line.

Contact

For inquiries, please contact 4point0@polymtl.ca

Ce contenu a été mis à jour le 2025-03-25 à 15 h 46 min.