The Geometry of Random Features

ثبت نشده
چکیده

We present an in-depth examination of the effectiveness of radial basis function kernel (beyond Gaussian) estimators based on orthogonal random feature maps. We show that orthogonal estimators outperform state-of-the-art mechanisms that use iid sampling under weak conditions for tails of the associated Fourier distributions. We prove that for the case of many dimensions, the superiority of the orthogonal transform over iid methods can be accurately measured by a property we define called the charm of the kernel, and that orthogonal random features provide optimal kernel estimators. Furthermore, we provide the first theoretical results which explain why orthogonal random features outperform unstructured on downstream tasks such as kernel ridge regression by showing that orthogonal random features provide kernel algorithms with better spectral properties than the previous state-of-the-art. Our results enable practitioners more generally to estimate the benefits from applying orthogonal transforms.

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Image Stitching of the Computed Radiology images Using a Pixel-Based Approach

In this paper, a method for automatic stitching of radiology images based on pixel features has been presented. In this method, according to the smooth texture of radiological images and in order to increase the number of the extracted features after quality enhancement of initial radiology images, 45 degree isotropic mask is applied to each radiology image to observe the image details. After t...

متن کامل

Sensitivity Analysis of the Effect of Pore Structure and Geometry on Petrophysical and Electrical Properties of Tight Media: Random Network Modeling

Several methodologies published in the literature can be used to construct realistic pore networks for simple rocks, whereas in complex pore geometry formations, as formed in tight reservoirs, such a construction still remains a challenge. A basic understanding of pore structure and topology is essential to overcome the challenges associated with the pore scale modeling of tight porous media. A...

متن کامل

Spatial Analysis in curved spaces with Non-Euclidean Geometry

The ultimate goal of spatial information, both as part of technology and as science, is to answer questions and issues related to space, place, and location. Therefore, geometry is widely used for description, storage, and analysis. Undoubtedly, one of the most essential features of spatial information is geometric features, and one of the most obvious types of analysis is the geometric type an...

متن کامل

Explaining the Role of Geometry and Imagination in Creation of Unique Architectures

A In this article the role of geometry and its interaction with imagination in creating the unique architectural will be explained. From long time ago philosophers and prominents have referred to important role that imagination and geometry have in the creation of universe. All over the universe, the emergence of geometry at beginning of creation is undeniable. Geometry with its manifestation a...

متن کامل

SIZE AND GEOMETRY OPTIMIZATION OF TRUSS STRUCTURES USING THE COMBINATION OF DNA COMPUTING ALGORITHM AND GENERALIZED CONVEX APPROXIMATION METHOD

In recent years, the optimization of truss structures has been considered due to their several applications and their simple structure and rapid analysis. DNA computing algorithm is a non-gradient-based method derived from numerical modeling of DNA-based computing performance by new computers with DNA memory known as molecular computers. DNA computing algorithm works based on collective intelli...

متن کامل

Parleda: a Library for Parallel Processing in Computational Geometry Applications

ParLeda is a software library that provides the basic primitives needed for parallel implementation of computational geometry applications. It can also be used in implementing a parallel application that uses geometric data structures. The parallel model that we use is based on a new heterogeneous parallel model named HBSP, which is based on BSP and is introduced here. ParLeda uses two main lib...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2017