Show NH [sic]: “data-to-paper” - autonomous stepwise LLM-driven research
data-to-paper is a framework for systematically navigating the power of AI to perform complete end-to-end scientific research, starting from raw data and concluding with comprehensive, transparent, and human-verifiable scientific papers
The example “research paper” was some useless fluff about diabetes, based off an existing data set (read: actual work produced by actual humans), and mad-libs.
The study identifies an inverse correlation between physical activity and fruit and vegetable intake with diabetes occurrence, while higher BMI is positively correlated
I’m too sleepy and statistics-impaired to check how nonsensical the regression “analysis” or findings are, so instead let’s check out the references (read: the actual humans who were plagarized to make this fluff)!
Reference #5
[5] T. Schnurr, Hermina Jakupovi, Germn D. Carrasquilla, L. ngquist, N. Grarup, T. Srensen, A. Tjnneland, K. Overvad, O. Pedersen, T. Hansen, and T. Kilpelinen. Obesity, unfavourable lifestyle and genetic risk of type 2 diabetes: a case-cohort study. Diabetologia, 63:1324–1332, 2020.
This incredibly managed to mangle all non-English alphabet names:
Hermina Jakupović, Germán D. Carrasquilla, Lars Ängquist, Thorkild I. A. Sørensen, Anne Tjønneland, Tuomas O. Kilpeläinen
I guess AI has an easier time advancing science than producing a PDF with non-ascii text in it