New metric space concept and quasi-isometry properties explored.
problem Exploring new metric spaces and quasi-isometry properties.
method Introducing (b,c)-metric and defining collapsing maps.
result Collapsing maps preserve quasi-isometry properties.
Study on energy of maps from K3 surface to flat orbifold.
problem Energy of maps from K3 surface to flat orbifold.
method Investigate Dirichlet energy of smooth maps and introduce an invariant.
result Ratio of energy to invariant converges to 1 for Foscolo's collapsing families.
This work builds the connection between the regularity theory of optimal transportation map, Monge-Ampère equation and GANs, which gives a theoretic understanding of the major drawbacks of GANs: convergence difficulty and mode collapse. According to the regularity theory of Monge-Ampère equation, if the support of the …
We present a new approach to simple homotopy theory of polyhedra using finite topological spaces. We define the concept of collapse of a finite space and prove that this new notion corresponds exactly to the concept of a simplicial collapse. More precisely, we show that a collapse of finite spaces induces a simplicial …
Variational autoencoders often collapse, showing latent variables are non-identifiable.
problem Posterior collapse in variational autoencoders due to non-identifiable latent variables.
method Proves latent variable non-identifiability causes posterior collapse. Proposes latent-identifiable models using Brenier maps and input convex neural networks.
result Latent-identifiable models resolve posterior collapse and provide meaningful representations.
Unified proof of smooth fibration theorems for collapsed manifolds.
problem Smooth fibration theorems for collapsed manifolds with Ricci curvature bounded below.
method Generalized Reifenberg condition and transformation technique for almost splitting maps.
result Unified proof of smooth fibration theorems in many previous works.
Study describes limits of non-collapsing K3 surfaces using algebraic data.
problem Understanding limits of non-collapsing polarized K3 surfaces.
method Explicit description via period mapping and algebro-geometric data.
result Bubbling limits depend solely on algebro-geometric data.
This study investigates abrupt learning dynamics in Transformers, revealing plateau formation and internal representation collapse.
problem Abrupt learning in Transformers, particularly during the loss plateau.
method Investigates mechanisms of abrupt learning in shallow Transformers, focusing on attention maps and hidden states.
result Reveals plateau formation, internal representation collapse, and strong repetition bias in outputs.
By Cheeger-Colding's almost splitting theorem, if a domain in a Ricci flat manifold is pointed-Gromov-Hausdorff close to a lower dimensional Euclidean domain, then there is a harmonic almost splitting map. We show that any eigenfunction of the Laplace operator is almost constant along the fibers of the almost splitting…
We exhibit families of Ricci-flat Kahler metrics on K3 surfaces which collapse to an interval, with Tian-Yau and Taub-NUT metrics occurring as bubbles. There is a corresponding continuous surjective map from the K3 surface to the interval, with regular fibers diffeomorphic to either 3-tori or Heisenberg nilmanifolds.
Compactness theory for biharmonic maps on degenerating Einstein manifolds.
problem Analyzing biharmonic maps on degenerating Einstein manifolds.
method Developed a compactness theory using asymptotic analysis over degenerating neck regions.
result Established a compactness theory for biharmonic maps with finitely many bubbles.
New model generates data without mode collapse or mixture.
problem Mode collapse and mode mixture in generative models.
method Inspired by AE-OT, improved sample generation algorithm.
result Free from mode collapse and mixture issues.
Study of splitting maps in Type I Ricci flows for understanding singular set structure.
problem Understanding the structure of the singular set in non-collapsed Ricci limit spaces.
method Construction and investigation of almost splitting maps on Ricci flows that are almost self-similar.
result Sharp splitting maps remain splitting maps at smaller scales under certain conditions.
Stability of Morse index for harmonic maps on degenerating surfaces analyzed.
problem Analyzing stability of Morse index for harmonic maps on degenerating Riemann surfaces.
method Analysis of second variation of energy, identification of conditions for upper semicontinuity, explicit contribution of geodesics.
result Sharper control of spectrum of Jacobi operator, explicit contribution of geodesic segments to Morse index.
Simplified proof of K3 surface period map surjectivity.
problem Surjectivity of period map on K3 surfaces.
method Utilizes hyperkähler geometry and collapsing techniques.
result Simple proof of Todorov's result on K3 surfaces.
We formulate a theory of pointed manifolds, accommodating both embeddings and Pontryagin-Thom collapse maps, so as to present a common generalization of Poincaré duality in topology and Koszul duality in En-algebra.
AGOP mechanism explains deep neural collapse in neural networks.
problem Explaining the rigid structure of data representations in deep neural networks.
method Introducing AGOP and Deep RFM to demonstrate DNC.
result AGOP mechanism causes deep neural collapse in neural networks.
iWGAN improves GANs by stabilizing training and preventing mode collapse.
problem Stable and effective training of GANs with mode collapse.
method iWGAN combines auto-encoders and WGANs using iterative primal dual optimization.
result iWGAN provides a clear stopping criterion and mitigates mode collapse.
We analyse the singularity formation of congruences of solutions of systems of second order PDEs via the construction of \emph{shape maps}. The trace of such maps represents a congruence volume whose collapse we study through an appropriate evolution equation, akin to Raychaudhuri's equation. We develop the necessary g…
Existence of harmonic maps near projections in hyperbolic spaces for large convex sets.
problem Existence of harmonic maps near projections in hyperbolic spaces.
method Weak quasi-isometry condition, non-collapsing property, harmonic measures.
result Existence of harmonic maps at finite distance from projections of large convex sets in hyperbolic spaces.
We study partial collapsing degeneration of Hamiltonian-perturbed Floer trajectories for an adiabatic ε-family and its reversal adiabatic gluing, as the prototype of the partial collapsing degeneration of 2-dimensional (perturbed) J-holomorphic maps to 1-dimensional gradient segments. We consider the …
Suppose a sequence Mj of Alexandrov spaces collapses to a space X with only weak singularities. Yamaguchi constructed a map fj:Mj→X called an almost Lipschitz submersion for large j. We prove that if Mj has a uniform positive lower bound for the volumes of spaces of directions, which is sufficiently la…
GANs mode collapse solved with Bures distance.
problem GANs mode collapse or mode dropping.
method Use Bures distance to match real and fake batch diversity in feature space.
result Diversity matching reduces mode collapse and improves sample quality.
It was recently proved that embedded solutions of Euclidean hypersurface flows with speeds given by concave (convex), degree one homogeneous functions of the Weingarten map are interior (exterior) non-collapsing. These results were subsequently extended to hypersurface flows in the sphere and hyperbolic space. In the f…
Similarity algebra extends algebraic structures with quantitative bounds.
problem Exact algebraic structures with strict axioms.
method Framework for approximate algebraic and Lie structures with ε-estimates. result Similarity structures converge to classical algebraic objects as εightarrow0. Deep generative models provide powerful tools for distributions over complicated manifolds, such as those of natural images. But many of these methods, including generative adversarial networks (GANs), can be difficult to train, in part because they are prone to mode collapse, which means that they characterize only a …
Proposes PSCs for UQ in deep nets without retraining.
problem Estimating uncertainty in deep nets with a single pass.
method Identifies sensitive, smooth intermediate layer, fits probabilistic model.
result PSCs achieve UQ and OOD detection performance matching existing methods.
For fibred boundary and fibred cusp metrics, Hausel, Hunsicker, and Mazzeo identified the space of L2 harmonic forms of fixed degree with the images of maps between intersection cohomology groups of an associated stratified space obtained by collapsing the fibres of the fibration at infinity onto its base. In the pr…
DVE uses GPs on DNN outputs to provide UQ without retraining.
problem Feature collapse in DNNs affects UQ methods.
method Deep Vecchia ensemble (DVE) of GPs on DNN hidden layers.
result Deterministic UQ possible in feature-collapsed DNNs.
Bayesian deep learning faces posterior collapse due to likelihood vs. prior competition.
problem Posterior collapse in Bayesian deep learning models.
method Identified competition between likelihood and prior regularization in a linear latent variable model.
result Posterior collapse is related to neural and dimensional collapse, suggesting a broader learning issue.
Encoder-decoder GANs architectures (e.g., BiGAN and ALI) seek to add an inference mechanism to the GANs setup, consisting of a small encoder deep net that maps data-points to their succinct encodings. The intuition is that being forced to train an encoder alongside the usual generator forces the system to learn meaning…
Proves weakly non-collapsed RCD spaces are strongly non-collapsed.
problem Proving the equivalence of weakly non-collapsed and strongly non-collapsed RCD spaces.
method Analyzes properties of RCD spaces and uses auxiliary results.
result Confirms conjecture about RCD spaces being strongly non-collapsed.
NullSpaceNet maps inputs to a joint-nullspace for clearer class separability.
problem Class separability and interpretability in image classification.
method NullSpaceNet maps inputs to a joint-nullspace, collapsing same-class inputs and separating different classes.
result NullSpaceNet achieves superior performance with reduced parameters and time.
Study on Neural Collapse limits in deep learning.
problem Understanding the limits of Neural Collapse in deep learning.
method Investigated Neural Collapse in the context of generalization and feature learning, refining conjectures and conducting experiments.
result Neural Collapse primarily occurs on the train set and not on the test set, suggesting it is an optimization phenomenon with unclear connections to generalization.
Study heat flow on collapsing K3 surfaces, handling conic singularities.
problem Analyzing heat flow on K3 surfaces as they collapse.
method Using semi-flat product approximations and conic-renormalized bilinear functionals.
result Heat operators converge to base Laplacian on regular locus.
The set of totally geodesic representatives of a homotopy class of maps from a compact Riemannian manifold M with nonnegative Ricci curvature into a complete Riemannian manifold N with no focal points is path-connected and, when nonempty, equal to the set of energy-minimizing maps in that class. When N is compact…
Nonuniform tubular neighborhoods of curves in Euclidean n-space are studied by using weighted distance functions and generalizing the normal exponential map. Different notions of injectivity radii are introduced to investigate singular but injective exponential maps. A generalization of the thickness formula is obtaine…
Ricci flow smooths locally collapsing manifolds with controlled curvature.
problem Locally collapsing manifolds with controlled Ricci curvature.
method Ricci flow for a definite period of time, detecting collapsing infranil fiber bundles.
result Topological conditions detect collapsing infranil fiber bundles.
Mathematical analysis shows annealing prevents mode collapse in Gaussian mixtures.
problem Mode collapse in variational inference for multimodal distributions.
method Analyzed annealing strategies for Gaussian mixtures, derived formulas, and tested on neural networks.
result Appropriately chosen annealing schemes can robustly prevent mode collapse.
The study characterizes and rules out collapsing in convex ancient mean curvature flow.
problem Characterizing and ruling out collapsing in convex ancient mean curvature flow.
method Characterization and counterexamples.
result Collapsing occurs if and only if the flow is asymptotic to at least one Grim hyperplane.
Despite excellent progress in recent years, mode collapse remains a major unsolved problem in generative adversarial networks (GANs).In this paper, we present spectral regularization for GANs (SR-GANs), a new and robust method for combating the mode collapse problem in GANs. Theoretical analysis shows that the optimal …
New method controls posterior collapse in VAEs without network architecture constraints.
problem Posterior collapse in VAEs reduces diversity of generated samples.
method Introduces Latent Reconstruction (LR) loss to control posterior collapse.
result Controls posterior collapse on various datasets without architectural constraints.
Study flat manifolds' collapsed limits as flat orbifolds.
problem Understanding collapsed limits of flat manifolds.
method Analyzing totally geodesic foliations and Gromov-Hausdorff limits.
result Identify collapsed limits as flat orbifolds and provide criteria for singularity.
Special Lagrangian submanifolds emerge from K3 surface collapse.
problem Understanding special Lagrangian submanifolds in K3 surface collapse.
method Lifting affine lines to degenerating sequences of special Lagrangian submanifolds.
result Constructing special Lagrangian two-spheres connecting Taub-NUT bubbles.
The torus cannot collapse to a segment under certain curvature conditions.
problem Impossibility of codimension-one collapse for surfaces of negative Euler characteristic.
method Analysis of Gaussian curvature and homology loops.
result The torus cannot collapse to a segment under similar conditions to surfaces of negative Euler characteristic.
Collapsibility is a combinatorial strengthening of contractibility. We relate this property to metric geometry by proving the collapsibility of any complex that is CAT(0) with a metric for which all vertex stars are convex. This strengthens and generalizes a result by Crowley. Further consequences of our work are: (1) …
Two-dimensional collapsed spaces with lower Ricci bounds are topological surfaces.
problem Topology of collapsed spaces with lower Ricci bounds
method Prove that collapsed spaces are topological surfaces
result Collapsed spaces are topological surfaces
Improved VAE models avoid posterior collapse in text modeling.
problem Posterior collapse in VAEs leads to poor data manifold parameterization.
method Coupled-VAE couples a VAE with a deterministic autoencoder to improve encoder and decoder parameterizations.
result Coupled-VAE consistently improves results in probability estimation and latent space richness.