An n-dimensional hybrid system embeds into a continuous vector field in m > 2n dimensions, enabling latent Neural ODEs with consistency losses to recover hybrid flows from time series.
Title resolution pending
6 Pith papers cite this work, alongside 220 external citations. Polarity classification is still indexing.
6
Pith papers citing it
220
external citations · OpenAlex
years
2026 6representative citing papers
Introduces coherence as a topological constraint on representations and the Coh objective to enforce geometric clustering for interpretability in neural networks.
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.
Time-delayed CROCKERs from persistent homology of simulated pedestrian positions separate corridor flow regimes by inflow parameters up to symmetry.
Reparameterization through coverings makes the KL term tractable in VAEs whose latent manifolds have non-trivial topology, demonstrated on a Klein bottle latent space.
Topological summaries of molecular clouds from the PHANGS-JWST catalogue correlate with physical characteristics of their galaxies.
citing papers explorer
-
Embedding Hybrid Systems into Continuous Latent Vector Fields
An n-dimensional hybrid system embeds into a continuous vector field in m > 2n dimensions, enabling latent Neural ODEs with consistency losses to recover hybrid flows from time series.
-
Learning Coherent Representations: A Topological Approach to Interpretability
Introduces coherence as a topological constraint on representations and the Coh objective to enforce geometric clustering for interpretability in neural networks.
-
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 study of holes: Topological analysis reveals crowd dynamics regimes in a bidirectional corridor scenario
Time-delayed CROCKERs from persistent homology of simulated pedestrian positions separate corridor flow regimes by inflow parameters up to symmetry.
-
Reparameterization through Coverings and Topological Weight Priors
Reparameterization through coverings makes the KL term tractable in VAEs whose latent manifolds have non-trivial topology, demonstrated on a Klein bottle latent space.
-
Topology of molecular clouds in the PHANGS-JWST catalogue in relation to the physical characteristics of their galaxies
Topological summaries of molecular clouds from the PHANGS-JWST catalogue correlate with physical characteristics of their galaxies.