Showing cs.LGShow all
3 papers · 1 filter
cs.LG2026
Towards Understanding the Shape of Representations in Protein Language Models
Kosio Beshkov, Anders Malthe-Sørenssen
While protein language models (PLMs) are one of the most promising avenues of research for future de novo protein design, the way in which they transform sequences to hidden repres…
cs.LG2026
A Quotient Homology Theory of Representation in Neural Networks
Kosio Beshkov
Previous research has proven that the set of maps implemented by neural networks with a ReLU activation function is identical to the set of piecewise linear continuous maps. Furthe…
cs.LG2024
A rank decomposition for the topological classification of neural representations
Kosio Beshkov, Gaute T. Einevoll
Neural networks can be thought of as applying a transformation to an input dataset. The way in which they change the topology of such a dataset often holds practical significance f…