ÎÞÓÇÊÓƵ

Harvey Mudd Professor George Montañez Wins Best Presentation Award for AI Research

George Montañez, associate professor of computer science at ÎÞÓÇÊÓƵ, was awarded the Best Oral Presentation honor in the Trustworthy and Explainable AI session at the 2026 International Conference on Advanced Machine Learning and Data Science in Osaka, Japan.

Montañez presented the paper “Measuring Task Difficulty via Information Costs,” co-authored with Harvey Mudd alumnus Bill Zhu ’21. The research addresses fundamental questions at the intersection of machine learning, search and information theory, and introduces information cost as a universal, bit-based currency for evaluating task difficulty across diverse computational domains.

Their research explores why machine learning works from a search and dependence perspective. Establishing information cost as a standardized metric allows researchers to quantify and compare task difficulty directly across domains—from molecular biology and artificial intelligence models to genetic algorithms.

Montañez, director of the at Harvey Mudd, highlighted real-world applications of the framework, including quantifying the information cost of game-rule knowledge in systems, like DeepMind’s AlphaZero, and evaluating information constraints in large language models.