SciIntBench: Measuring LLM Compliance with Research Integrity Norms Under Adversarial Framing
Research overviewAlmene De Meran Meguimtsop, Maria Leonor Pacheco, Daniel E. Acuna
Science is
our subject.
We study how science works.
And build tools to help it work better.
The Science of Science & Computational Discovery Lab at the University of Colorado Boulder.
Explore our researchBetter questions.
Better science.
From the integrity of a single figure to the structure of the scientific community, we examine the systems that shape discovery.
Can we trust
the scientific record?
Research integrity
Detecting image manipulation, evaluating questionable journals, and examining integrity in AI-generated research.
How can AI help
science move forward?
Computational discovery
Developing methods for peer review, knowledge extraction, and research recommendation.
Who gets to
shape discovery?
Science of science
Understanding how mentorship, funding, collaboration, and inequality influence the research ecosystem.
Ideas in the making.
All publicationsREM-CTX: Automated Peer Review via Reinforcement Learning with Auxiliary Context
Research overviewPawin Taechoyotin, Daniel E. Acuna
Dissecting the gender divide: authorship and acknowledgment in scientific publications
Research overviewKeigo Kusumegi, Daniel E. Acuna, Yukie Sano
Estimating the predictability of questionable open-access journals
Research overviewHan Zhuang, Lizhen Liang, Daniel E. Acuna
Curiosity is
a team effort.
We bring computational methods and a shared interest in science to questions that cross disciplinary boundaries.
Meet the people
From the lab.
All news & events- December 2026Daniel serves as an Area Chair for the AI for Science workshop at NeurIPS 2026
- November 2026Daniel serves as a reviewer for the 4th Workshop on Artificial Intelligence for Scientific Publications (WASP 2026) at AACL-IJCNLP 2026
- September 28, 2026Daniel will deliver a keynote, AI vs. AI: Guarding the Scientific Record Against Fabrication, Manipulation, and Machine-Authored Content, in the Ethics & Governance in AI for Science track of AI4Sci Korea 2026, Seoul, South Korea (September 28 – October 1, 2026)
- August 2026Dissecting the gender divide: authorship and acknowledgment in scientific publications (Keigo Kusumegi, Daniel, and Yukie Sano) is published in Scientometrics
Join Us
Information about doctoral study, student research, and visiting appointments in the lab.
Department of Computer Science
University of Colorado Boulder
Help the next idea take shape.
Support our mission to understand and improve science.