arXiv research

New method for accurately predicting linear dynamical systems.

Langevin Dynamics fails to sample from mixture distributions efficiently.

Generalizes LD approach to create minimal surfaces and self-shrinkers.

FA-LD algorithm improves uncertainty quantification and mean predictions in federated learning.

LD-EnSF speeds up data assimilation with sparse observations.

The paper quantizes concatenated noisy vectors to a common cluster center, improving performance over naive methods.

We develop a novel method for training of GANs for unsupervised and class conditional generation of images, called Linear Discriminant GAN (LD-GAN). The discriminator of an LD-GAN is trained to maximize the linear separability between distributions of hidden representations of generated and targeted samples, while the …

Improved diffusion bridge sampling with rKL-LD loss.

Linear Dynamical System (LDS) is an elegant mathematical framework for modeling and learning multivariate time series. However, in general, it is difficult to set the dimension of its hidden state space. A small number of hidden states may not be able to model the complexities of a time series, while a large number of …

Algorithm learns multiple LDS models from unlabeled short trajectories.

RQMC improves QMC by providing practical error bounds for financial applications.

We propose local distributional smoothness (LDS), a new notion of smoothness for statistical model that can be used as a regularization term to promote the smoothness of the model distribution. We named the LDS based regularization as virtual adversarial training (VAT). The LDS of a model at an input datapoint is defin…

We consider a setting, where the output of a linear dynamical system (LDS) is, with an unknown but fixed probability, replaced by noise. There, we present a robust method for the prediction of the outputs of the LDS and identification of the samples of noise, and prove guarantees on its statistical performance. One app…

Paper proposes a new algorithm combining gradient descent and Langevin dynamics.

dCMF models evolving patterns in multiway data with temporal dynamics.

A local deformation property for uniform embeddings in metric manifolds (LD) is formulated and its behaviour is studied in a formal view point. It is shown that any metric manifold with a geometric group action, typical metric spaces (Euclidean space, hyperbolic space and cylinders) and for κ\leq 0 the κ-cone ends over…

Recently, the technique of local updates is a powerful tool in centralized settings to improve communication efficiency via periodical communication. For decentralized settings, it is still unclear how to efficiently combine local updates and decentralized communication. In this work, we propose an algorithm named as L…

We propose a novel multilinear dynamical system (MLDS) in a transform domain, named $\mathcal{L}$-MLDS, to model tensor time series. With transformations applied to a tensor data, the latent multidimensional correlations among the frontal slices are built, and thus resulting in the computational independence in the tra…

Langevin dynamics (LD) has been proven to be a powerful technique for optimizing a non-convex objective as an efficient algorithm to find local minima while eventually visiting a global minimum on longer time-scales. LD is based on the first-order Langevin diffusion which is reversible in time. We study two variants th…

Levenshtein VAE prevents posterior collapse in text generation models.

Method combines LD and Fermat Distance for neural network uncertainty.

SNF combines stochastic and deterministic steps to sample complex distributions.

Personalization of cardiac models involves the optimization of organ tissue properties that vary spatially over the non-Euclidean geometry model of the heart. To represent the high-dimensional (HD) unknown of tissue properties, most existing works rely on a low-dimensional (LD) partitioning of the geometrical model. Wh…

Tensor-EM method learns MoLDS from complex, noisy data.

RLD improves combinatorial optimization by avoiding local minima.

Efficient algorithm predicts unknown linear systems with long-term memory.

In recent years, manifold learning has become increasingly popular as a tool for performing non-linear dimensionality reduction. This has led to the development of numerous algorithms of varying degrees of complexity that aim to recover man ifold geometry using either local or global features of the data. Building on t…

New algorithm recovers mixture means even with many outliers.

This paper proposes a methodology for host-based anomaly detection using a semi-supervised algorithm namely one-class classifier combined with a PCA-based feature extraction technique called Eigentraces on system call trace data. The one-class classification is based on generating a set of artificial data using a refer…

TiDE uses MLP for fast, simple long-term time-series forecasting.

Optimal sample complexity for learning DDAGs from noisy data.

Proposes TgNN-LD to improve neural network effectiveness and efficiency.

A new method de-randomizes MCMC dynamics using the Stein operator.

In earlier work of NK new closed embedded smooth minimal surfaces in the round three-sphere $\mathbb{S}^3(1)$ were constructed, each resembling two parallel copies of the equatorial two-sphere $\mathbb{S}^2_{eq}$ joined by small catenoidal bridges, with the catenoidal bridges concentrating along two parallel circles, o…

Gaussian process (GP) audio source separation is a time-domain approach that circumvents the inherent phase approximation issue of spectrogram based methods. Furthermore, through its kernel, GPs elegantly incorporate prior knowledge about the sources into the separation model. Despite these compelling advantages, the c…

Study on tradeoffs between mistakes and ERM oracle calls in online and transductive learning.

New algorithm learns stable LDSs with lower error and better control performance.

Spectral method speeds fitting of binary time series models.

We revisit the Bayesian online inference problems for the linear dynamic systems (LDS) under non- Gaussian environment. The noises can naturally be non-Gaussian (skewed and/or heavy tailed) or to accommodate spurious observations, noises can be modeled as heavy tailed. However, at the cost of such noise robustness, the…

FA-HMC improves Bayesian federated learning with rigorous guarantees.

New self-shrinkers with multiple ends constructed by stacking planes.

The paper tackles fairness in forecasting and learning linear dynamical systems.

RAPTOR-GEN accelerates digital twin development for biopharmaceuticals.

Study on order book dynamics with uniform catastrophes, explaining volatility and trends.

New algorithms improve sampling from complex distributions.

LaPSRL achieves optimal regret for isoperimetric RL distributions.

Improves generative models by using heavy-tailed noise in score matching.

We present a unified framework to analyze the global convergence of Langevin dynamics based algorithms for nonconvex finite-sum optimization with $n$ component functions. At the core of our analysis is a direct analysis of the ergodicity of the numerical approximations to Langevin dynamics, which leads to faster conver…