Výsledky vyhledávání - "Dean, Nathaniel"

Report

Novel Deep Neural Network Classifier Characterization Metrics with Applications to Dataless Evaluation

The mainstream AI community has seen a rise in large-scale open-source classifiers, often pre-trained on vast datasets and tested on standard benchmarks; however, users facing diverse needs and limited, expensive test data may be overwhelmed by avail

Externí odkaz: http://arxiv.org/abs/2407.13000

Zobrazit plný text záznamu

Report

Minimizing Chebyshev Prototype Risk Magically Mitigates the Perils of Overfitting

Autor: Dean, Nathaniel, Sarkar, Dilip

Overparameterized deep neural networks (DNNs), if not sufficiently regularized, are susceptible to overfitting their training examples and not generalizing well to test data. To discourage overfitting, researchers have developed multicomponent loss f

Externí odkaz: http://arxiv.org/abs/2404.07083

Zobrazit plný text záznamu

Report

Fantastic DNN Classifiers and How to Identify them without Data

Autor: Dean, Nathaniel, Sarkar, Dilip

Current algorithms and architecture can create excellent DNN classifier models from example data. In general, larger training datasets result in better model estimations, which improve test performance. Existing methods for predicting generalization

Externí odkaz: http://arxiv.org/abs/2305.15563

Zobrazit plný text záznamu

Report

A Perturbation Resistant Transformation and Classification System for Deep Neural Networks

Autor: Dean, Nathaniel, Sarkar, Dilip

Deep convolutional neural networks accurately classify a diverse range of natural images, but may be easily deceived when designed, imperceptible perturbations are embedded in the images. In this paper, we design a multi-pronged training, input trans

Externí odkaz: http://arxiv.org/abs/2208.11839

Zobrazit plný text záznamu

Report

Coprime and prime labelings of graphs

Autor: Berliner, Adam H., Dean, Nathaniel, Hook, Jonelle, Marr, Alison, Mbirika, Aba, McBee, Cayla D.

Publikováno v: Journal of Integer Sequences, Vol. 19 (2016), Article 16.5.8

A coprime labeling of a simple graph of order $n$ is a labeling in which adjacent vertices are given relatively prime labels, and a graph is prime if the labels used can be taken to be the first $n$ positive integers. In this paper, we consider when

Externí odkaz: http://arxiv.org/abs/1604.07698

Zobrazit plný text záznamu

Report

Iteration Index of a Zero Forcing Set in a Graph

Autor: Chilakamarri, Kiran B., Dean, Nathaniel, Kang, Cong X., Yi, Eunjeong

Publikováno v: Bull. Inst. Combin. Appl. Vol. 64 (2012) pp. 57-72

Let each vertex of a graph G = (V(G), E(G)) be given one of two colors, say, "black" and "white". Let Z denote the (initial) set of black vertices of G. The color-change rule converts the color of a vertex from white to black if the white vertex is t

Externí odkaz: http://arxiv.org/abs/1105.1492

Zobrazit plný text záznamu