arXiv research

This paper explains how model invariance improves generalization using data transformations.

Clever sampling methods can be used to improve the handling of big data and increase its usefulness. The subject of this study is remote sensing, specifically airborne laser scanning point clouds representing different classes of ground cover. The aim is to derive a supervised learning model for the classification usin…

Several approximate policy iteration schemes without value functions, which focus on policy representation using classifiers and address policy learning as a supervised learning problem, have been proposed recently. Finding good policies with such methods requires not only an appropriate classifier, but also reliable e…

New bounds for private learning of high-dimensional Gaussian distributions.

Generative adversarial approach for satellite image time series land cover classification.

Bob predicts a future observation based on a sample of size one. Alice can draw a sample of any size before issuing her prediction. How much better can she do than Bob? Perhaps surprisingly, under a large class of loss functions, which we refer to as the Cover-Hart family, the best Alice can do is to halve Bob's risk. …

Study shows offline RL with partial coverage and weak function classes is possible.

This paper reviews various sampling methods from statistics and machine learning.

This paper presents a Bayesian optimization method with exponential convergence without the need of auxiliary optimization and without the delta-cover sampling. Most Bayesian optimization methods require auxiliary optimization: an additional non-convex global optimization problem, which can be time-consuming and hard t…

In the panoply of pattern classification techniques, few enjoy the intuitive appeal and simplicity of the nearest neighbor rule: given a set of samples in some metric domain space whose value under some function is known, we estimate the function anywhere in the domain by giving the value of the nearest sample per the …

New insights show coverage conditions are crucial for efficient online reinforcement learning.

Two algorithms for interpreting and boosting tree-based models using rule covering.

Study shows a central limit theorem for random coverings of manifolds with nilpotent groups.

We classify all closed, aspherical Riemannian manifolds M whose universal cover has indiscrete isometry group. One sample application is the theorem that any such M with word-hyperbolic fundamental group must be isometric to a negatively curved, locally symmetric manifold. Another application is the classification of a…

We analyze a new Markov chain model for better sampling and optimization.

New SGD covering technique yields dimension-independent generalization bounds.

We consider model-free reinforcement learning for infinite-horizon discounted Markov Decision Processes (MDPs) with a continuous state space and unknown transition kernel, when only a single sample path under an arbitrary policy of the system is available. We consider the Nearest Neighbor Q-Learning (NNQL) algorithm to…

The paper tightens bounds on covering numbers for deep ReLU networks.

Proper regularization is critical for speeding up training, improving generalization performance, and learning compact models that are cost efficient. We propose and analyze regularized gradient descent algorithms for learning shallow neural networks. Our framework is general and covers weight-sharing (convolutional ne…

Study shows private learning of mixtures of Gaussians is possible with polynomial samples.

Study on covering probability of random balls in bounded open sets.

Generative adversarial networks (GANs) are a powerful framework for generative tasks. However, they are difficult to train and tend to miss modes of the true data generation process. Although GANs can learn a rich representation of the covered modes of the data in their latent space, the framework misses an inverse map…

Methods for prediction and tolerance intervals in non-normal models.

Deep convolutional neural networks can be highly vulnerable to small perturbations of their inputs, potentially a major issue or limitation on system robustness when using deep networks as classifiers. In this paper we propose a low-cost method to explore marginal sample data near trained classifier decision boundaries…

Deep learning models perform variably across continents/seasons in land cover mapping.

This paper proposes an online tree-based Bayesian approach for reinforcement learning. For inference, we employ a generalised context tree model. This defines a distribution on multivariate Gaussian piecewise-linear models, which can be updated in closed form. The tree structure itself is constructed using the cover tr…

The paper improves transformer generalization bounds using rank-dependent covering number bounds.

Paper introduces a new framework to improve sample efficiency in POMDPs learning.

Active Learning (AL) is a learning task that requires learners interactively query the labels of the sampled unlabeled instances to minimize the training outputs with human supervisions. In theoretical study, learners approximate the version space which covers all possible classification hypothesis into a bounded conve…

New property ensures neural networks generalize well with limited data.

Adaptive allocation with constraints using Thompson sampling.

The paper tackles learning smooth distance functions using query-based methods.

Standard adversarial training involves two agents, namely a generator and a discriminator, playing a mini-max game. However, even if the players converge to an equilibrium, the generator may only recover a part of the target data distribution, in a situation commonly referred to as mode collapse. In this work, we prese…

Transfer learning across different reinforcement learning (RL) tasks is becoming an increasingly valuable area of research. We consider a goal-based multi-task RL framework and mechanisms by which previously solved tasks can reduce sample complexity and regret when the agent is faced with a new task. Specifically, we i…

This is a detailed tutorial paper which explains the Principal Component Analysis (PCA), Supervised PCA (SPCA), kernel PCA, and kernel SPCA. We start with projection, PCA with eigen-decomposition, PCA with one and multiple projection directions, properties of the projection matrix, reconstruction error minimization, an…

New RL approach uses resets to learn complex MDPs efficiently.

Unified framework for data-free sampling using Wasserstein gradient flows.

Long horizon reinforcement learning is as hard as short horizon learning.

Improved DDPMs achieve high log-likelihoods and sample quality with fewer passes.

We provide a differentially private algorithm for hypothesis selection. Given samples from an unknown probability distribution $P$ and a set of $m$ probability distributions $\mathcal{H}$, the goal is to output, in a $\varepsilon$-differentially private manner, a distribution from $\mathcal{H}$ whose total variation di…

The paper proves generalization bounds and stopping rules for self-selected data in reciprocal learning.

Study covers of surfaces, showing types and properties.

Unified analysis of generalization and sample complexity for semi-supervised domain adaptation.

A {\em solvable} cover of a graph is a regular cover whose covering transformation group is solvable. In this paper, we show that a solvable cover of a graph can be decomposed into layers of abelian covers, and also, a lift of a given automorphism of the base graph of a solvable cover can be decomposed into layers of l…

Survey of mathematical foundations for reinforcement learning.

Unified method for MMD variance estimation improves accuracy and computational efficiency.

After showing that a covering space of surface bundles over $S^1$ factors as a `covering of fibers' followed by a `power covering', we prove that, for torus bundles, power coverings do not lower Heegaard genus, and that fiber coverings lower the genus only in special cases.

Federated Q-learning achieves linear speedup with heterogeneity, improving sample complexity.