arXiv research

Parsimonious neural networks discover interpretable physical laws from data.

Method selects most useful network model for various tasks.

The parsimonious Gaussian mixture models, which exploit an eigenvalue decomposition of the group covariance matrices of the Gaussian mixture, have shown their success in particular in cluster analysis. Their estimation is in general performed by maximum likelihood estimation and has also been considered from a parametr…

Bayesian method reconstructs hidden higher-order interactions from network data.

Investigates deep hedging under rough volatility models.

Combining Bayesian nonparametrics and a forward model selection strategy, we construct parsimonious Bayesian deep networks (PBDNs) that infer capacity-regularized network architectures from the data and require neither cross-validation nor fine-tuning when training the model. One of the two essential components of a PB…

Kolmogorov-Arnold Networks offer improved interpretability and parsimony in science tasks.

This paper proposes a parsimoniously time varying parameter vector autoregressive model (with exogenous variables, VARX) and studies the properties of the Lasso and adaptive Lasso as estimators of this model. The parameters of the model are assumed to follow parsimonious random walks, where parsimony stems from the ass…

We consider the problem of modeling multivariate time series with parsimonious dynamical models which can be represented as sparse dynamic Bayesian networks with few latent nodes. This structure translates into a sparse plus low rank model. In this paper, we propose a Gaussian regression approach to identify such a mod…

A parsimonious model reduces over-parameterization in skewed matrix variate mixtures.

Enhances KANs for accuracy and interpretability with multi-exit architecture.

Solving for adversarial examples with projected gradient descent has been demonstrated to be highly effective in fooling the neural network based classifiers. However, in the black-box setting, the attacker is limited only to the query access to the network and solving for a successful adversarial example becomes much …

A family of parsimonious shifted asymmetric Laplace mixture models is introduced. We extend the mixture of factor analyzers model to the shifted asymmetric Laplace distribution. Imposing constraints on the constitute parts of the resulting decomposed component scale matrices leads to a family of parsimonious models. An…

A family of parsimonious Gaussian cluster-weighted models is presented. This family concerns a multivariate extension to cluster-weighted modelling that can account for correlations between multivariate responses. Parsimony is attained by constraining parts of an eigen-decomposition imposed on the component covariance …

We introduce a methodology to construct parsimonious probabilistic models. This method makes use of Information Filtering Networks to produce a robust estimate of the global sparse inverse covariance from a simple sum of local inverse covariances computed on small sub-parts of the network. Being based on local and low-…

Proposes a model for identifying edges in low-rank dynamical networks.

We propose an elementary model to price European physical delivery swaptions in multicurve setting with a simple exact closed formula. The proposed model is very parsimonious: it is a three-parameter multicurve extension of the two-parameter Hull-White (1990) model. The model allows also to obtain simple formulas for a…

A new model captures financial asset returns' tail behaviors and outperforms GARCH family.

The use of artificial neural networks as models of chaotic dynamics has been rapidly expanding. Still, a theoretical understanding of how neural networks learn chaos is lacking. Here, we employ a geometric perspective to show that neural networks can efficiently model chaotic dynamics by becoming structurally chaotic t…

We study "active" decision making over sensor networks where the sensors' sequential probing actions are actively chosen by continuously learning from past observations. We consider two network settings: with and without central coordination. In the first case, the network nodes interact with each other through a centr…

A neural network approach solves dynamic portfolio optimization without dynamic programming.

New framework for network regression models accounting for community structure.

In this paper we consider sparse and identifiable linear latent variable (factor) and linear Bayesian network models for parsimonious analysis of multivariate data. We propose a computationally efficient method for joint parameter and model inference, and model comparison. It consists of a fully Bayesian hierarchy for …

Extracts coarse-grained PDEs from microscopic simulations.

Improved SINDy autoencoder for identifying noisy dynamical systems.

Machine Learning algorithms are increasingly being used in recent years due to their flexibility in model fitting and increased predictive performance. However, the complexity of the models makes them hard for the data analyst to interpret the results and explain them without additional tools. This has led to much rese…

Finite mixtures of regression models offer a flexible framework for investigating heterogeneity in data with functional dependencies. These models can be conveniently used for unsupervised learning on data with clear regression relationships. We extend such models by imposing an eigen-decomposition on the multivariate …

New GMM models fit high-dimensional data with fewer parameters.

We consider the problem of non-parametric regression with a potentially large number of covariates. We propose a convex, penalized estimation framework that is particularly well-suited for high-dimensional sparse additive models. The proposed approach combines appealing features of finite basis representation and smoot…

Bayesian context trees capture complex dependencies in categorical sequences.

Proposes a non-convex optimization method for a parsimonious weighted naive Bayes classifier.

New neural network models for complex functional data analysis.

Proposes a new framework for predicting stock market movements using sparse neural architectures.

Novel method combines neural network features with survival models for ICU infections.

A new method reduces high-dimensional data's impact on CWMs using TSNE.

Let $G$ be a nonabelian, simple group with a nontrivial conjugacy class $C \subseteq G$. Let $K$ be a diagram of an oriented knot in $S^3$, thought of as computational input. We show that for each such $G$ and $C$, the problem of counting homomorphisms $π_1(S^3\setminus K) \to G$ that send meridians of $K$ to $C$ is al…

GIT-Net uses neural networks to approximate PDE operators efficiently.

New neural networks model complex phenomena with fewer parameters.

New method for learning multidimensional CDFs using Archimedean copulas.

Parsimonious Dynamic Mode Decomposition selects sparse modes robustly.

Proposes a new approach to approximate maximum likelihood for complex models.

Optimal AFs minimize RFR test error and sensitivity.

Novel estimation methods improve MAR model accuracy for high-dimensional time series.

We investigate the detectability of modules in large networks when the number of modules is not known in advance. We employ the minimum description length (MDL) principle which seeks to minimize the total amount of information required to describe the network, and avoid overfitting. According to this criterion, we obta…

MDL principle aids in learning neural network-based causal structures.

Path regularization reveals convex optimization in deep ReLU networks.

Recent years have demonstrated that using random feature maps can significantly decrease the training and testing times of kernel-based algorithms without significantly lowering their accuracy. Regrettably, because random features are target-agnostic, typically thousands of such features are necessary to achieve accept…

Overparameterisation can limit the benefits of curriculum learning in neural networks.