Paper presented at ICCV 2019.
This paper targets the task with discrete and periodic
class labels (e.g., pose/orientation estimation) in the context of deep learning. The commonly used cross-entropy or
regression loss is not well matched to this problem as they
ignore the periodic nature of the labels and the class similarity, or assume labels are continuous value. We propose to
incorporate inter-class correlations in a Wasserstein training framework by pre-defining (i.e., using arc length of a
circle) or adaptively learning the ground metric. We extend
the ground metric as a linear, convex or concave increasing
function w.r.t. arc length from an optimization perspective.
We also propose to construct the conservative target labels
which model the inlier and outlier noises using a wrapped
unimodal-uniform mixture distribution. Unlike the one-hot
setting, the conservative label makes the computation of
Wasserstein distance more challenging. We systematically
conclude the practical closed-form solution of Wasserstein
distance for pose data with either one-hot or conservative
target label. We evaluate our method on head, body, vehicle and 3D object pose benchmarks with exhaustive ablation studies. The Wasserstein loss obtaining superior performance over the current methods, especially using convex mapping function for ground metric, conservative label,
and closed-form solution.
Xiaofeng Liu, Yang Zou, Tong Che, Peng Ding, Ping Jia, Jane You, B.V.K. Vijaya Kumar
This is a working paper template developed by the Interdisciplinary Laboratory of Computational Social Science at the University of Maryland, College Park.
This is a template for a Lab report following an IEEE paper. Modified by Francisco Tovar after Michael Sheel original document.
This document will be used for EEET2493 class.
This document serves both as a template and a guideline for your report. You can replace headings and body text with your content. Before you start writing, however, we recommend that you carefully consider the instructions in this document.
En los últimos años se ha visto un auge en el uso de los sistemas de bases de datos NoSQL y junto a ello se ha popularizado la idea de aplicaciones de Persistencia Políglota. Esta consiste en que gracias a la gran variedad y cantidad de datos, y los diversos servicios que pueden dar las aplicaciones hoy en día, es probable que un único tipo de sistema de almacenamiento no sea capaz de cubrir de forma eficiente todas las necesidades de la aplicación. En este articulo se dará una idea general de las Aplicaciones de Persistencia Políglota dando información acerca de su funcionamiento, arquitectura y motivación; y ademas se hablara específicamente de como aplicar la Persistencia Políglota con MongoDB y Neo4j.
Palabras Clave: NoSQL, Persistencia Políglota, MongoDB, Neo4j, Neo4j Doc Manager