2 papers
cs.LG2026
Learning Coherent Representations: A Topological Approach to Interpretability
Sigurd Gaukstad, Melvin Vaupel, Valdemar Kargård Olsen +2
Deep neural networks learn representations where individual features often lack interpretable meaning; a single neuron may activate for scattered, unrelated inputs. We introduce co…
math.AT2025
Lifting Cocycles: From Heuristic to Theory
Sigurd Gaukstad, Mathias Karsrud Nordal, Marius Thaule
The circular coordinates algorithm, a key tool in topological data analysis, relies on a theoretically unvalidated lifting step to convert cocycles from a prime field to integer co…