Research
On-device research index

arXiv research

A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.

168,695 papers · 148 categories

Trend · papers per month

4048081,2121,616 · Jun 202019922001200920172026
48 results for Curvature Learning

New method efficiently learns positive-definite curvature for neural nets.

problem Efficiently learn positive-definite curvature for neural net training.
method Spectral-factorized positive-definite curvature learning approach.
result Efficiently applies arbitrary matrix roots and generic curvature learning.

This paper reviews discrete curvature models for geometric data analysis.

problem Capturing intrinsic geometric structure in diverse data representations.
method Comprehensive review of discrete curvature models from Riemannian and metric geometry perspectives.
result Systematic pipeline for curvature-driven data analysis and learning.

Spectro-Riemannian Graph Neural Networks integrate spectral and curvature signals for better graph representation learning.

problem Enhance graph representation learning by leveraging spectral and curvature signals.
method Proposes Spectro-Riemannian Graph Neural Networks (CUSP) that combines spectral and curvature insights.
result Empirical evaluation shows CUSP outperforms state-of-the-art models by up to 5.3%.

Deep learning predicts curvature of 2D interfaces in level-set method.

problem Estimating curvature in level-set method for complex interfaces.
method Deep learning using feed-forward neural networks trained on synthetic data.
result Deep learning models approximate curvature with comparable precision to traditional methods.

New method approximates curvature from symmetries in deep networks.

problem Hard to approximate curvature in large deep networks.
method Analytically averaging over group actions that leave the loss invariant to construct structured Hessian approximations.
result Structured Hessian approximations from single gradients can be estimated, stored, and inverted.

CurvSSL improves SSL by aligning local manifold curvature.

problem Improving self-supervised learning by capturing local manifold geometry.
method CurvSSL augments Barlow Twins with a curvature-based regularizer to align and decorrelate embeddings across augmentations.
result Curvature-regularized SSL yields competitive or improved linear evaluation performance.

Warm-up improves training by adapting learning rate based on curvature.

problem Improving training efficiency in deep learning models.
method Introducing a curvature condition to explain warm-up, and showing empirically that it leads to faster convergence.
result Adapting learning rate based on curvature condition naturally induces warm-up-like schedule, leading to faster convergence.

Study shows how discrete graph curvature relates to manifold curvature.

problem Relating discrete graph curvature to intrinsic manifold curvature.
method Continuum limits of Ollivier's Ricci curvature on data clouds.
result Random geometric graphs inherit global curvature properties of manifolds.

Curvature penalties improve interpretability of KANs without sacrificing accuracy.

problem Pathologically high-curvature oscillations in KANs activations make them hard to interpret.
method Derived a curvature penalty and proved an upper bound on model curvature.
result KANs with curvature penalties achieve substantially smoother activations while maintaining accuracy.

Curvature regularization prevents distortion in graph embeddings.

problem Graph topology patterns distort in Euclidean space, making detection difficult.
method Proposes curvature regularization to enforce flatness in embedding manifolds.
result Significant improvements in five embedding methods on open graph datasets.

Graph networks struggle with multi-task learning due to varying property loss surface curvatures.

problem Graph networks underperform in multi-task learning for crystal and molecule properties.
method Assessed curvature of property loss surfaces via spectral properties of Hessians, matrix-free using randomized numerical linear algebra.
result Varying curvature of property loss surfaces explains graph networks' multi-task learning inefficiency.

We propose meta-curvature (MC), a framework to learn curvature information for better generalization and fast model adaptation. MC expands on the model-agnostic meta-learner (MAML) by learning to transform the gradients in the inner optimization such that the transformed gradients achieve better generalization performa…

2019-02-09abs ↗pdf ↗

CAMEL enhances manifold embedding and learning with curvature metrics.

problem High-dimensional data classification, dimension reduction, and visualization.
method CAMEL uses a Riemannian manifold with curvature metrics for enhanced expressibility and interpretability.
result CAMEL outperforms state-of-the-art methods on high-dimensional datasets.

New neural network solves Nirenberg problem for curvature on sphere.

problem Prescribing Gaussian curvature on S2S^2 for metrics conformal to the round metric.
method Mesh-free physics-informed neural network (PINN) that directly parametrises the conformal factor.
result Neural network achieves very low losses for realisable curvatures, distinguishing them from non-realisable ones.

This thesis explores Ollivier-Ricci curvature in graphs and manifolds, with applications to graph neural networks.

problem Understanding curvature in metric spaces and graphs.
method Combines optimal transport theory, Riemannian manifolds, and graph theory to define and analyze Ollivier-Ricci curvature.
result Extensions of Ollivier-Ricci curvature to directed graphs and applications in network science.

We give a relationship that yields an effective geometric way of evaluating mean curvature of surfaces. The approach is reminiscent of the Gauss's contour based evaluation of intrinsic curvature. The presented formula may have a number of potential applications including estimating the normal vector and mean curvature …

2011-08-07abs ↗pdf ↗

New bounds for neural networks on curved manifolds improve generalization.

problem Existing generalization theories fail to account for non-Euclidean manifold structures.
method Derive covering number bounds incorporating manifold-specific properties like curvature.
result Sharp Rademacher complexity bounds for neural networks on compact manifolds.

CWGD measures gradient diversity weighted by curvature, improving SGD convergence.

problem Gradient noise in high-curvature directions is underestimated by standard methods.
method CWGD weights gradient diversity by the inverse square root of the Hessian.
result CWGD-Cosine reduces optimization error by up to 20% compared to standard cosine annealing.

New framework to understand and exploit curvature in deep learning loss landscapes.

problem Understanding and optimizing the loss landscape in deep learning models.
method New conceptual framework and techniques to estimate and exploit curvature of expected loss changes.
result Alice algorithm optimizes training by incorporating curvature terms and step bounds.

Euclidean geometry has historically been the typical "workhorse" for machine learning applications due to its power and simplicity. However, it has recently been shown that geometric spaces with constant non-zero curvature improve representations and performance on a variety of data types and downstream tasks. Conseque…

2019-11-19abs ↗pdf ↗

This paper speeds up mean curvature computation for high-dimensional data.

problem Efficiently computing mean curvature in high-dimensional datasets.
method Two contributions: algebraic identity and truncated SVD approximation.
result Mean curvature computation reduced from O(m4)O(m^4) to O(k2m+kmp2)O(k^2 m + k m p^2).

RicciNets prunes neural networks by removing edges of low importance based on Ricci curvature, reducing FLOPs by 35%.

problem Pruning neural networks to reduce computational load and improve efficiency.
method RicciNets uses Ricci curvature to prune edges of low importance in a randomly wired neural network, reducing FLOPs.
result Reduction of almost 35% in FLOPs with no performance degradation.

The purpose of these notes is to provide an introduction to those who want to learn more about translating solitons for the mean curvature flow in R3\mathbb{R}^3, particularly those which are complete graphs over domains in R2\mathbb{R}^2. In this paper we describe a full classification of complete translating graphs i…

2019-01-25abs ↗pdf ↗

Stochastic gradient algorithms have been the main focus of large-scale learning problems and they led to important successes in machine learning. The convergence of SGD depends on the careful choice of learning rate and the amount of the noise in stochastic estimates of the gradients. In this paper, we propose a new ad…

2014-12-23abs ↗pdf ↗

New research shows hyperbolic embeddings are useful for global consistency tasks in graphs.

problem The usefulness of hyperbolic representations in graph learning tasks.
method Computed hyperbolic embeddings for node classification and link prediction tasks, addressing optimization issues at zero curvature.
result Hyperbolic embeddings are more effective for tasks requiring global consistency, while Euclidean models are superior for other tasks.

Paper investigates hardness of learning neural networks under manifold hypothesis.

problem Hardness of learning neural networks under the manifold hypothesis.
method Extending proofs of hardness in the SQ and cryptographic settings to the geometric setting.
result Learning is hard under input manifolds of bounded curvature but learnable with additional assumptions on manifold volume.

In topological data analysis, persistent homology is used to study the "shape of data". Persistent homology computations are completely characterized by a set of intervals called a bar code. It is often said that the long intervals represent the "topological signal" and the short intervals represent "noise". We give ev…

2019-05-30abs ↗pdf ↗

Improved continual learning for neural networks with BN layers using K-FAC extension.

problem Continual learning challenges in neural networks with BN layers.
method Extended K-FAC method to account for inter-example relations, weight merging, and reparameterization for BN layers; proposed weight merging and reparameterization for BN layers; proposed method to select hyperparameters without source task data.
result Better performance in continual learning tasks with BN layers compared to baselines.

Consider a sample of nn points taken i.i.d from a submanifold ΣΣ of Euclidean space. We show that there is a way to estimate the Ricci curvature of ΣΣ with respect to the induced metric from the sample. Our method is grounded in the notions of Carré du Champ for diffusion semi-groups, the theory of Empirical process…

2014-10-13abs ↗pdf ↗

The paper reveals that baselines significantly impact RL algorithms' convergence.

problem Understanding the true impact of baselines on policy optimization.
method Theoretical analysis of bandit and RL problems, focusing on natural policy gradient and EXP3.
result Baselines can determine algorithm convergence, contradicting traditional optimization theory.

The question of how to incorporate curvature information in stochastic approximation methods is challenging. The direct application of classical quasi- Newton updating techniques for deterministic optimization leads to noisy curvature estimates that have harmful effects on the robustness of the iteration. In this paper…

2014-01-27abs ↗pdf ↗