arXiv research

Proposes a copula-based model for multi-view clustering with directional dependency.

Estimates mean of random vector with near-optimal error in all directions.

Neural network pruning is an important step in design process of efficient neural networks for edge devices with limited computational power. Pruning is a form of knowledge transfer from the weights of the original network to a smaller target subnetwork. We propose a new method for compute-constrained structured channe…

New approach uses Gaussian processes to learn and track complex systems with guaranteed accuracy.

Anisotropy of a space naturally leads to direction dependent electromagnetic tensors and electromagnetic potentials. Starting from this idea and using variational approaches and exterior derivative formalism, we extend some of the classical equations of electromagnetism to anisotropic (Finslerian) spaces. The results d…

New algorithms reduce regret in neural logistic bandits.

The paper validates statistical models for groundwater data.

This paper presents a new open source Python framework for causal discovery from observational data and domain background knowledge, aimed at causal graph and causal mechanism modeling. The 'cdt' package implements the end-to-end approach, recovering the direct dependencies (the skeleton of the causal graph) and the ca…

Survey of methods to recover CI graphs from feature relationships.

Recent work on adversarial attack has shown that Projected Gradient Descent (PGD) Adversary is a universal first-order adversary, and the classifier adversarially trained by PGD is robust against a wide range of first-order attacks. It is worth noting that the original objective of an attack/defense model relies on a d…

We briefly review some basic concepts of parallel displacement in Finsler geometry. In general relativity, the parallel translation of objects along the congruence of the fundamental observer corresponds to the evolution in time. By dropping the quadratic restriction on the measurement of an infinitesimal distance, the…

Paper uses relaxation techniques to find optimal brokerage fees with private signals.

We investigate approaches to regularisation during fine-tuning of deep neural networks. First we provide a neural network generalisation bound based on Rademacher complexity that uses the distance the weights have moved from their initial values. This bound has no direct dependence on the number of weights and compares…

Deep learning has exhibited superior performance for various tasks, especially for high-dimensional datasets, such as images. To understand this property, we investigate the approximation and estimation ability of deep learning on anisotropic Besov spaces. The anisotropic Besov space is characterized by direction-depen…

The paper examines how market trade values and volumes affect price autocorrelation.

Improved confidence bounds for linear logistic model with applications to bandits.

New framework detects directional influence in multivariate time series.

The paper studies Möbius energy gradient of helix pairs and finds limiting behavior as coiling ratio increases.

Anisotropic obstacle problems and Stefan problem studied with evolving surfaces.

The conventional formal tool to detect effects of the financial persistence is in terms of the Hurst exponent. A typical corresponding result is that its value comes out close to 0.5, as characteristic for geometric Brownian motion, with at most small departures from this value in either direction depending on the mark…

"Mixed Data" comprising a large number of heterogeneous variables (e.g. count, binary, continuous, skewed continuous, among other data types) are prevalent in varied areas such as genomics and proteomics, imaging genetics, national security, social networking, and Internet advertising. There have been limited efforts a…

We introduce a new approach to functional causal modeling from observational data, called Causal Generative Neural Networks (CGNN). CGNN leverages the power of neural networks to learn a generative model of the joint distribution of the observed variables, by minimizing the Maximum Mean Discrepancy between generated an…

The paper analyzes convergence of Langevin dynamics with time-dependent metrics.

Direct approach for handling contextual bandits with latent state dynamics.

The geometry of causal diamonds or Alexandrov open sets whose initial and final events $p$ and $q$ respectively have a proper-time separation $τ$ small compared with the curvature scale is a universal. The corrections from flat space are given as a power series in $τ$ whose coefficients involve the curvature at the cen…

Unsupervised ensemble learning has long been an interesting yet challenging problem that comes to prominence in recent years with the increasing demand of crowdsourcing in various applications. In this paper, we propose a novel method-- unsupervised ensemble learning via Ising model approximation (unElisa) that combine…

ManifoldFlow relaxes fixed-spectrum Stiefel layers to learn a positive spectrum.

Regulating causal effects through averaged constraints fails to enforce conditional independence.

In this study, we present a simple stochastic order-book model for investors' swarm behaviors seen in the continuous double auction mechanism, which is employed by major global exchanges. Our study shows a characteristic called "fat tail" is seen in the data obtained from our model that incorporates the investors' swar…

Paper proposes an algorithm to learn DAGs with indirect dependencies.

EPSTE: A geometric token and deep learning approach to estimating transfer entropy in neuroimaging time series

Sharp rates found for learning with dependent data, avoiding sample size deflation.

The edge structure of the graph defining an undirected graphical model describes precisely the structure of dependence between the variables in the graph. In many applications, the dependence structure is unknown and it is desirable to learn it from data, often because it is a preliminary step to be able to ascertain c…

New method for robust financial portfolio analysis.

Neural architecture search (NAS) relies on a good controller to generate better architectures or predict the accuracy of given architectures. However, training the controller requires both abundant and high-quality pairs of architectures and their accuracy, while it is costly to evaluate an architecture and obtain its …

Proposes a new adversarial model to avoid accuracy vs. adversarial accuracy tradeoff.

Selective classification can worsen accuracy disparities between groups.

Calibrated ensembles improve both ID and OOD accuracy in distribution shift.

Paper monitors DNN accuracy to enhance trustworthiness.

Optimizes glmnet configuration for better accuracy and efficiency.

TRUST improves tree models' accuracy while maintaining interpretability.

FADE framework improves fairness and accuracy in ensemble learning.

Paper explores tradeoff between standard and robust accuracy for latent models.

New methods show robustness and accuracy can coexist.

Improved forecast accuracy for Knitwear by 20% using adaptive AI/ML model.

Proposes an accuracy-preserving calibration method for DNNs.

The paper studies and mitigates accuracy disparity in regression models.

New work shows limits of certifying neural network robustness.