The National Science Foundation Division of Undergraduate Education awarded Texas State University $291,134 on November 1, 2026, to advance methods for analyzing mathematical proof language under the STEM Education program (CFDA 47.076).
The project, titled Advancing Methods for Analyzing Proof Language (AMPL), develops AI-based approaches to identify and characterize language conventions in formal mathematical proofs. The recipient will build two new educational corpora—one from undergraduate mathematics textbooks and one from student proof attempts collected across undergraduate classrooms and online platforms—alongside an existing corpus of mathematician publications. Analysis will employ qualitative human coding, weak supervision, and semantic embeddings to identify linguistic patterns the project designates as proof-language markers. The work identifies divergences between student conventions and expert conventions, enabling more targeted instruction on proof-writing expectations. Project outputs will be openly available to researchers and educators to support improved access to proof-based mathematics instruction.
Work is performed in San Marcos, Texas, with a period of performance from November 1, 2026, through October 31, 2028. The assistance type is a project grant.