Researchers have conducted a study to determine the entropy of the Ukrainian language, a measure of its unpredictability. Using a method similar to Claude Shannon's 1951 experiment, 184 volunteers predicted characters in Ukrainian sentences. The study established an upper bound for Ukrainian language entropy at approximately 1.201 bits per character. The findings were compared against the performance of current Large Language Models, and the methods and code were made publicly available. AI
Summary written by gemini-2.5-flash-lite from 2 sources. How we write summaries →
IMPACT Provides a benchmark for Ukrainian language complexity, aiding LLM development and evaluation for the language.
RANK_REASON Academic paper published on arXiv detailing a new experiment and findings.