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.
PUMA augments models to remove unique data points without performance loss.
In our previous works, we proposed a physically-inspired rule to organize the data points into an in-tree (IT) structure, in which some undesired edges are allowed to occur. By removing those undesired or redundant edges, this IT structure is divided into several separate parts, each representing one cluster. In this w…
Minimal surfaces and curves can have singularities removed by isotopy.
Unified framework for removing unwanted information from machine learning models.
New framework for attributing online marketing touchpoints.
We call a singularity of a presymplectic form removable in its graph if its graph extends to a smooth Dirac structure over the singularity. An example for this is the symplectic form of a magnetic monopole. A criterion for the removability of singularities is given in terms of regularizing functions for pure spinor…
Ray Singer torsion is a numerical invariant associated with a compact Riemannian manifold equipped with a flat bundle and a Hermitian structure on this bundle. In this note we show how one can remove the dependence on the Riemannian metric and on the Hermitian structure with the help of a base point and of an Euler str…
Anisotropic min-max theory constructs stable minimal surfaces in 3-manifolds.
A statistical framework for removing unwanted data domains in machine learning.
A new method detects outliers in dirty data using a leave-out strategy.
Paper presents a new way to estimate model changes without full model evaluation.
We assess cluster stability by trimming extreme points and tracking data range reduction.
Good data stewardship requires removal of data at the request of the data's owner. This raises the question if and how a trained machine-learning model, which implicitly stores information about its training data, should be affected by such a removal request. Is it possible to "remove" data from a machine-learning mode…
In machine learning, statistics, econometrics and statistical physics, cross-validation (CV) is used asa standard approach in quantifying the generalisation performance of a statistical model. A directapplication of CV in time-series leads to the loss of serial correlations, a requirement of preserving anynon-stationar…
We clarify the status of log-periodicity associated with speculative bubbles preceding financial crashes. In particular, we address Feigenbaum's [2001] criticism and show how it can be rebuked. Feigenbaum's main result is as follows: ``the hypothesis that the log-periodic component is present in the data cannot be reje…
We use methods of complex analysis to extend the bundle structure across a removable point-singularity in a Sasakian three-manifold.
The study calculates the index distribution of Brownian loops in various geometrical settings.
Selective removal of data subsets can efficiently unlearn unwanted distributions.
Bayesian inference forgetting framework removes influence of single data points.
We consider harmonic maps into pseudo-Riemannian manifolds. We show the removability of isolated singularities for continuous maps, i.e. that any continuous map from an open subset of R^m into a pseudo-Riemannian manifold which is two times continuously differentiable and harmonic everywhere outside an isolated point i…
Paper improves DP-ERM for binary linear classification with large-margin subsets.
Gradient ascent method successfully removes specific data points from neural networks without retraining.
Following on from ``Hyperbolic Plateau problems'' (by the same author), we provide a complete geometric description of solutions to the Plateau problem when is a compact Riemann surface with a finite number of points removed.
The importance of training robust neural network grows as 3D data is increasingly utilized in deep learning for vision tasks in robotics, drone control, and autonomous driving. One commonly used 3D data type is 3D point clouds, which describe shape information. We examine the problem of creating robust models from the …
Let be a closed Riemann surface of genus with one point removed. In this paper, we identify those point-pushing pseudo-Anosov maps on that preserve at least one bi-infinite geodesic in the curve complex.
Traditional text classifiers are limited to predicting over a fixed set of labels. However, in many real-world applications the label set is frequently changing. For example, in intent classification, new intents may be added over time while others are removed. We propose to address the problem of dynamic text classifi…
New method estimates mean from noisy data with few outliers.
New SVM margin bound improves generalization in machine learning.
We study isolated singularities of two dimensional Yang-Mills-Higgs fields defined on a fiber bundle, where the fiber space is a compact Riemannian manifold and the structure group is a compact connected Lie group. In general the singularity can not be removed due to possibly non-vanishing limit holonomy around the sin…
Computes fundamental groups of restricted configuration spaces.
Influence functions estimate the effect of removing a training point on a model without the need to retrain. They are based on a first-order Taylor approximation that is guaranteed to be accurate for sufficiently small changes to the model, and so are commonly used to study the effect of individual points in large data…
D3M debiases models by selectively removing problematic examples.
Proposes a new method to unlearn from specific data points in conformal predictors.
This paper proposes a novel Gaussian process approach to fault removal in time-series data. Fault removal does not delete the faulty signal data but, instead, massages the fault from the data. We assume that only one fault occurs at any one time and model the signal by two separate non-parametric Gaussian process model…
Data selection methods, such as active learning and core-set selection, are useful tools for machine learning on large datasets. However, they can be prohibitively expensive to apply in deep learning because they depend on feature representations that need to be learned. In this work, we show that we can greatly improv…
Let X be a compact 2-manifold with nonempty boundary dX and let f: (X, dX) --> (X, dX) be a boundary-preserving map. Denote by MF_d[f] the minimum number of fixed point among all boundary-preserving maps that are homotopic through boundary-preserving maps to f. The relative Nielsen number N_d(f) is the sum of the numbe…
New method stabilizes model selection with theoretical guarantees.
Proposes DeGLIF to denoise graph data for label noise robustness.
We pursue the analogy of a framed flow category with the flow data of a Morse function. In classical Morse theory, Morse functions can sometimes be locally altered and simplified by the Morse moves. These moves include the Whitney trick which removes two oppositely framed flowlines between critical points of adjacent i…
DaRE forests enable efficient data deletion from random forests.
Local gaps in Ricci shrinkers depend only on dimension.
We construct a sequence of compact embedded minimal disks in the unit ball in Euclidean 3-space whose boundaries are in the boundary of the ball and where the curvatures blow up at every point of a line segment of the vertical axis, extending from the origin. We further study the transversal structure of the minimal li…
Cluster analysis and outlier detection are strongly coupled tasks in data mining area. Cluster structure can be easily destroyed by few outliers; on the contrary, outliers are defined by the concept of cluster, which are recognized as the points belonging to none of the clusters. Unfortunately, most existing studies do…
Study eigenfunctions of Laplacian on sphere with even point removals.
We study graphs of positive extrinsic curvature with a non-removable isolated singularity in 3-dimensional warped product spaces, and describe their behavior at the singularity in several natural situations. We use Monge-Ampère equations to give a classification of the surfaces in 3-dimensional space forms which are em…
Squint bound improved by removing term.
This paper studies the regularity of constrained Willmore immersions into locally around both "regular" points and around branch points, where the immersive nature of the map degenerates. We develop local asymptotic expansions for the immersion, its first, and its second derivatives, given in terms of resi…
We demonstrate, theoretically and empirically, that adversarial robustness can significantly benefit from semisupervised learning. Theoretically, we revisit the simple Gaussian model of Schmidt et al. that shows a sample complexity gap between standard and robust classification. We prove that unlabeled data bridges thi…