Online Archive of University of Virginia Scholarship
Persistent Convolution: A Topological Approach to Formal AI Alignment Testing25 views
Author
Ashoff, Tyler, Statistics - Graduate School of Arts and Sciences, University of Virginia
Advisors
Kafadar, Karen, AS-Statistics (STAT), University of Virginia
Rodu, Jordan, AS-Statistics (STAT), University of Virginia
Abstract
This work compares topological representations of AI models’ embeddings to formally test the alignment of their semantic spaces against interpretable baselines. Persistence landscapes and silhouettes offer a way to characterize global and local structure that focuses on connectivity and shape rather than scale. Since these representations are agnostic to the embedding dimensionality, different model architectures can be compared consistently. This also allows for intuitive low-dimensional relationships between concepts or formal knowledge structures like ontologies to be used as baselines for alignment tests. The results show that these methods can effectively be used for model selection and training by tracking semantic structure as it is induced, optimized, or degraded.
Degree
PHD (Doctor of Philosophy)
Keywords
AI Alignment; Topological Data Analysis; Semantic Structure; Mechanistic Interpretability; Persistent Homology
Ashoff, Tyler. Persistent Convolution: A Topological Approach to Formal AI Alignment Testing. University of Virginia, Statistics - Graduate School of Arts and Sciences, PHD (Doctor of Philosophy), 2026-05-01, https://doi.org/10.18130/8k9j-9k42.