Kirill Solovev

About Me

I am a researcher at the IDea_Lab of the University of Graz. My main research interests lie in applying frontier machine learning methods to solve problems from various disciplines and hopefully make the world a tad nicer. My current work for the Horizon Europe project VALPOP is a pipeline that allows extracting relationship networks from millions of multilingual news articles. My colleagues and I have also released a sub-second-queryable dataset of 1.3 billion articles, and plan to study how narratives and political concepts move and evolve across cultures, languages, and time.

For most of the papers I've worked on, you'll find the code on my Codeberg. Support open science!

In my free time I try to provide learning help that I struggled with. Check out Blattspiel (FOSS, obviously), a sight-reading trainer that makes practice a bit more fun.

Research Interests:

• Computational Humanities
• Narrative & Conceptual Change
• Network Analysis & Prosopography
• Natural Language Processing & LLMs
• Information Extraction & Knowledge Graphs
• Multimodal & Computer Vision
• Political Communication & Online Discourse
• Open Science
• Computational Social Science