arXiv research

This paper surveys various methods for dimensionality reduction and nearest neighbor search.

The approximation of nonlinear kernels via linear feature maps has recently gained interest due to their applications in reducing the training and testing time of kernel-based learning algorithms. Current random projection methods avoid the curse of dimensionality by embedding the nonlinear feature space into a low dim…

Develops an oblique projection technique to approximate a foliation for non-normal dynamics.

New method learns nonlinear projections for reduced-order modeling of complex dynamical systems.

DiffSlack learns neural networks with nonlinear constraints via learnable slack variables.

The paper introduces a fast algorithm for learning and forecasting nonlinear dynamics from noisy time series data.

A method models nonlinear dynamics from data using barycentric coordinates and memory.

Proves a theorem for Assouad dimension with applications to distance sets and radial projections.

We study the use of "sign $α$-stable random projections" (where $0<α\leq 2$) for building basic data processing tools in the context of large-scale machine learning applications (e.g., classification, regression, clustering, and near-neighbor search). After the processing by sign stable random projections, the inner pr…

In this paper we investigate for further symmetry properties of the nonlinear fin equations of the general form $u_t=(E(u)u_x)_x + h(x)u$ rather than recent works on these equations. At first, we study the projective (fiber-preserving) symmetry to show that equations of the above class can not be reduced to linear equa…

Develops optimal low-dimensional approximations to high-dimensional SDEs.

Study tackles nonlinear factor models with unknown monotone links from incomplete and noisy data.

In this paper we study the problem of recovering a structured but unknown parameter ${\bfθ}^*$ from $n$ nonlinear observations of the form $y_i=f(\langle {\bf{x}}_i,{\bfθ}^*\rangle)$ for $i=1,2,\ldots,n$. We develop a framework for characterizing time-data tradeoffs for a variety of parameter estimation algorithms when…

Proposes MFPC for cross-manifold clustering.

PGD algorithms solve nonlinear inverse problems with generative priors using noisy measurements.

K-means is a classical clustering algorithm with wide applications. However, soft K-means, or fuzzy c-means at m=1, remains unsolved since 1981. To address this challenging open problem, we propose a novel clustering model, i.e. Probabilistic K-Means (PKM), which is also a nonlinear programming model constrained on lin…

In this paper, we study a class of Finsler metrics composed by a Riemann metric $α=\sqrt{a_{ij}(x)y^i y^j}$ and a $1$-form $β=b_i(x)y^i$ called general ($α$, $β$)-metrics. We classify those projectively flat when $α$ is projectively flat. By solving the corresponding nonlinear PDEs, the metrics in this class are totall…

This study develops a NURBS-based method for conformal surface flattening without singularities.

Properties of the Cauchy-Riemann-Fueter equation for maps between quaternionic manifolds are studied. Spaces of solutions in case of maps from a K3-surface to the cotangent bundle of a complex projective space are computed. A relationship between harmonic spinors of a generalized nonlinear Dirac operator and solutions …

A novel optimisation framework through quadratic nonlinear projection is introduced for credit portfolio when the portfolio risk is measured by Conditional Value-at-Risk (CVaR). The whole optimisation procedure to search toward the optimal portfolio state is conducted by a series of single-step optimisations under the …

In this paper we introduce a projection method for the space of probability distributions based on the differential geometric approach to statistics. This method is based on a direct L2 metric as opposed to the usual Hellinger distance and the related Fisher Information metric. We explain how this apparatus can be used…

The paper examines how nonlinear transformations affect ridge sets in manifold learning.

Using nonlinear pde techniques, we construct a new family of globally smooth tt* structures. This includes tt* structures associated to the (orbifold) quantum cohomology of a finite number of complex projective spaces and weighted projective spaces. The existence of such "magical solutions" of the tt* equations, namely…

We propose a novel adaptive learning algorithm based on iterative orthogonal projections in the Cartesian product of multiple reproducing kernel Hilbert spaces (RKHSs). The task is estimating/tracking nonlinear functions which are supposed to contain multiple components such as (i) linear and nonlinear components, (ii)…

The goal of this paper is to provide a geometric framework for analyzing the uniform decay properties of solutions to the Teukolsky equation in the fully nonlinear setting of perturbations of Kerr. It contains the first nonlinear version of the Chandrasekhar transformation introduced in the linearized setting in \cite{…

FEALM learns features for better nonlinear DR of hidden patterns.

It has been reported repeatedly that discriminative learning of distance metric boosts the pattern recognition performance. A weak point of ITML-based methods is that the distance threshold for similarity/dissimilarity constraints must be determined manually and it is sensitive to generalization performance, although t…

A new DDR framework learns low-dimensional data representations using dynamical systems.

Ancient solutions arise in the study of Ricci flow singularities. Motivated by the work of Fateev on 3-dimensional ancient solutions we construct high dimensional ancient solutions to Ricci flow on spheres and complex projective spaces as well as the twistor spaces over a compact quaternion-Kahler manifold. Differing f…

A new deep neural network tackles nonlinear functional regression with improved dimensionality reduction.

Bayesian approach to portfolio selection reduces pessimism in frequent trading.

For many years, a combination of principal component analysis (PCA) and independent component analysis (ICA) has been used for blind source separation (BSS). However, it remains unclear why these linear methods work well with real-world data that involve nonlinear source mixtures. This work theoretically validates that…

Geometric theory of projection heads in self-supervised learning.

Unified theory and debiasing framework for random oblique projections in high dimensions.

New method converts LVAs into linear projections for better understanding of complex models.

In this paper, we study locally projectively flat Finsler metrics with constant flag curvature ${\bf K}$. We prove those are totally determined by their behaviors at the origin by solving some nonlinear PDEs. The classifications when ${\bf K}=0$, ${\bf K}=-1$ and ${\bf K} =1$ are given respectively in an algebraic way.…

Geometrically characterizes virtual nonlinear nonholonomic constraints using symplectic methods.

The method of random projections has become a standard tool for machine learning, data mining, and search with massive data at Web scale. The effective use of random projections requires efficient coding schemes for quantizing (real-valued) projected data into integers. In this paper, we focus on a simple 2-bit coding …

Kernel approximation via nonlinear random feature maps is widely used in speeding up kernel machines. There are two main challenges for the conventional kernel approximation methods. First, before performing kernel approximation, a good kernel has to be chosen. Picking a good kernel is a very challenging problem in its…

In this paper we investigate a nonlinear generalization of the Black-Scholes equation for pricing American style call options in which the volatility term may depend on the underlying asset price and the Gamma of the option. We propose a numerical method for pricing American style call options by means of transformatio…

New insights into continual learning for deep models, showing convergence issues but local linear solutions.

Paper presents a fast and adaptive filter for SI suppression in full-duplex transceivers.

DeepRSCN models nonlinear systems using stochastic configurations.

Enhances RSCNs with hybrid regularization for nonlinear dynamics.

Researchers found new functions for spherical clothoids using special functions.

Develops MGQDA for multi-group classification with theoretical guarantees and practical applications.

Connections between integration along hypersufaces, Radon transforms, and neural networks are exploited to highlight an integral geometric mathematical interpretation of neural networks. By analyzing the properties of neural networks as operators on probability distributions for observed data, we show that the distribu…

Kernel discriminant analysis uses nonlinear embeddings to improve classification.