SMIXAE is a new mixture-of-autoencoders architecture that learns multidimensional manifolds directly from transformer activations, recovering known structures and identifying novel ones in Gemma 2 2B and 9B models.
Nicolas Bonneel, Julien Rabin, Gabriel Peyr ´e, and Hanspeter Pfister
4 Pith papers cite this work. Polarity classification is still indexing.
4
Pith papers citing it
years
2026 4verdicts
UNVERDICTED 4representative citing papers
A parameterized family of tensor products on persistence modules produces Künneth short exact sequences and universal coefficient theorems usable for persistent homology of filtered CW complexes and product spaces.
Random slicing for subsampling combined with Nadaraya-Watson smoothing enables faster and improved persistence-based topological optimization of point clouds in 2D and 3D.
A standardized pipeline converts time series to graphs, computes persistence diagrams, and extracts features that classify UCR benchmarks, with diffusion distance outperforming shortest-path metrics and performance varying by graph type.
citing papers explorer
-
SMIXAE: Towards Unsupervised Manifold Discovery in Language Models
SMIXAE is a new mixture-of-autoencoders architecture that learns multidimensional manifolds directly from transformer activations, recovering known structures and identifying novel ones in Gemma 2 2B and 9B models.
-
A continuum of K\"unneth theorems for persistence modules
A parameterized family of tensor products on persistence modules produces Künneth short exact sequences and universal coefficient theorems usable for persistent homology of filtered CW complexes and product spaces.
-
Towards Scalable Persistence-Based Topological Optimization
Random slicing for subsampling combined with Nadaraya-Watson smoothing enables faster and improved persistence-based topological optimization of point clouds in 2D and 3D.
-
Persistent Homology of Time Series through Complex Networks
A standardized pipeline converts time series to graphs, computes persistence diagrams, and extracts features that classify UCR benchmarks, with diffusion distance outperforming shortest-path metrics and performance varying by graph type.