Rasoul Norouzi

Department of Methodology and Statistics

Tilburg University

I am a phd researcher in the Department of Methodology and Statistics at Tilburg University. I am supervised by Dr. Caspar J. van Lissa, Dr. Bennett Kleinberg, and Professor Jeroen Vermunt.

My research focuses on theory development and refinement in the social sciences using natural language processing methods. I combine large language models for information extraction with graph-based methods for analyzing causal claims and supporting theory building.

Research interests: natural language processing, causal text mining, recommender systems, graph neural networks, and AI-powered research tools.

Lab affiliations: I am a member of the Insight Lab (transparent and reproducible social science research), the CPCM Lab (computational approaches to human cognition and behavior), and the Meta-Research Lab (studying and improving scientific research practices).

TMSR – eScience Project (Advisor & Contributor). Developing a Python CLI package to batch-download papers via DOI by integrating multiple provider APIs into one unified workflow for systematic reviews. I advised on project scope and requirements, and contributed to the evaluation design and methodological review. Stack: Python (CLI), API integration. GitHub · Project Page

Tools I work with include Python, JavaScript, React, NumPy, scikit-learn, PyTorch, Keras, Flask, and Node.js.

news

Jun 10, 2026	Presenting a poster at the SRSM Annual Meeting 2026 in Chania, Crete, Greece (June 10–12). Poster title: From Causal Extraction to Concept Harmonization: A Computational Pipeline for Automated Qualitative Research Synthesis. The work presents a BERT-based multitask model for parsing causal relations from social science publications (matching human inter-rater reliability, Krippendorff’s alpha = 0.80) and a multidimensional benchmark for concept harmonization methods. Together, these studies move toward a full pipeline for automated research synthesis.
Apr 24, 2026	Announcing a new Pre-Conference Workshop on Text Mining Systematic Reviews on September 30, 2026, organized in collaboration with the eScience Center NL and the Open Science Community Tilburg. Thanks to eScience Center’s expertise, participants will learn four main skills: (1) automatically downloading published papers in bulk for systematic review and text mining (2) constructing theories based on causal claims in published papers, (3) harmonizing social science concepts (e.g., recognizing that “sleep disorder” and “insomnia” refer to the same concept), and (4) knowledge extraction of relevant information from published papers using specialized models with known validity and reliability — no generative AI, only evidence-based, deterministic LLM pipelines.
Apr 20, 2026	Started supervising Amirali Rezazadeh (jointly with Bennett Kleinberg) on his traineeship. The goal is to (i) learn the fundamentals of classical natural language processing, (ii) implement methods in Python or R, and (iii) use these insights to read key papers of modern NLP applied to psychological research and synthesise these readings in a report.