http://www.cs.toronto.edu/~hinton/absps/transauto6.pdf

The artificial neural networks that are used to recognize shapes typically use one or more layers of learned feature detectors that produce scalar outputs. By contrast, the computer vision community uses complicated, hand-engineered features, like SIFT [6], that producea whole vector of outputs including an explicit representation of the pose of the feature. We show how neural networks can be used to learn features that output a whole vector of instantiation parameters and we argue that this is a much more promising way of dealing with variations in position,orientation, scale and lighting than the methods currently employed in the neural networks community. It is also more promising than the hand-engineered features currently used in computer vision because it provides an efficient way of adapting the features to the domain.

Current methods for recognizing objects in images perform poorly and use methods that are intellectually unsatisfying. Some of the best computer vision systems use histograms of oriented gradients as “visual words” and model the spatial distribution of these elements using a crude spatial pyramid. Such methods can recognize objects correctly without knowing exactly where they are – an ability that is used to diagnose brain damage in humans. The best artifical neural net-works [4, 5, 10] use hand-coded weight-sharing schemes to reduce the number of free parameters and they achieve local translational invariance by subsampling the activities of local pools of translated replicas of the same kernel. This method of dealing with the changes in images caused by changes in view point is much better than no method at all, but it is clearly incapable of dealing with recognition tasks, such as facial identity recognition, that require knowledge of the precise spatial relationships between high-level parts like a nose and a mouth.After several stages of subsampling in a convolutional net, high-level feature shave a lot of uncertainty in their poses. This is generally regarded as a desireable property because it amounts to invariance to pose over some limited range,but it makes it impossible to compute precise spatial relationships.

【This paper argues that convolutional neural networks are misguided in what they are trying to achieve.】

【capsules】

This paper argues that convolutional neural networks are misguided in what they are trying to achieve. Instead of aiming for viewpoint invariance in the activities of “neurons” that use a single scalar output to summarize the activities of a local pool of replicated feature detectors, artifical neural networks should use local “capsules” that perform some quite complicated internal computations on their inputs and then encapsulate the results of these computations into a small vector of highly informative outputs. Each capsule learns to recognize an implicitly defined visual entity over a limited domain of viewing conditions and deformations and it outputs both the probability that the entity is present within its limited domain and a set of “instantiation parameters” that may include the precise pose, lighting and deformation of the visual entity relative to an implicitly defined canonical version of that entity. When the capsule is working properly, the probability of the visual entity being present is locally invariant – it does not change as the entity moves over the manifold of possible appearances within the limited domain covered by the capsule. The instantiation parameters, however, are “equivariant” – as the viewing conditions change and the entity moves over the appearance manifold, the instantiation parameters change by a corresponding amount because they are representing the intrinsic coordinates of the entity on the appearance manifold.

Transforming Auto-encoders的更多相关文章

  1. [Python] 机器学习库资料汇总

    声明:以下内容转载自平行宇宙. Python在科学计算领域,有两个重要的扩展模块:Numpy和Scipy.其中Numpy是一个用python实现的科学计算包.包括: 一个强大的N维数组对象Array: ...

  2. python数据挖掘领域工具包

    原文:http://qxde01.blog.163.com/blog/static/67335744201368101922991/ Python在科学计算领域,有两个重要的扩展模块:Numpy和Sc ...

  3. Theano3.1-练习之初步介绍

    来自 http://deeplearning.net/tutorial/,虽然比较老了,不过觉得想系统的学习theano,所以需要从python--numpy--theano的顺序学习.这里的资料都很 ...

  4. [resource]Python机器学习库

    reference: http://qxde01.blog.163.com/blog/static/67335744201368101922991/ Python在科学计算领域,有两个重要的扩展模块: ...

  5. 机器学习——深度学习(Deep Learning)

    Deep Learning是机器学习中一个非常接近AI的领域,其动机在于建立.模拟人脑进行分析学习的神经网络,近期研究了机器学习中一些深度学习的相关知识,本文给出一些非常实用的资料和心得. Key W ...

  6. Deep Learning Tutorial - Classifying MNIST digits using Logistic Regression

    Deep Learning Tutorial 由 Montreal大学的LISA实验室所作,基于Theano的深度学习材料.Theano是一个python库,使得写深度模型更容易些,也可以在GPU上训 ...

  7. [转]Python机器学习工具箱

    原文在这里  Python在科学计算领域,有两个重要的扩展模块:Numpy和Scipy.其中Numpy是一个用python实现的科学计算包.包括: 一个强大的N维数组对象Array: 比较成熟的(广播 ...

  8. 深度学习材料:从感知机到深度网络A Deep Learning Tutorial: From Perceptrons to Deep Networks

    In recent years, there’s been a resurgence in the field of Artificial Intelligence. It’s spread beyo ...

  9. Deep Learning(4)

    四.拓展学习推荐 Deep Learning 经典阅读材料: The monograph or review paper Learning Deep Architectures for AI (Fou ...

  10. 深度学习教程Deep Learning Tutorials

    Deep Learning Tutorials Deep Learning is a new area of Machine Learning research, which has been int ...

随机推荐

  1. 在4418平台上如何配置GPIO口的状态

    硬件 ------------------------------------------------------------------------------------------------- ...

  2. 浅谈MVP设计模式

    最近公司在做一个医疗项目,使用WinForm界面作为客户端交互界面.在整个客户端解决方案中.使用了MVP模式实现.由于之前没有接触过该设计模式,所以在项目完成到某个阶段时,将使用MVP的体会写在博客里 ...

  3. AC日记——爱改名的小融2 codevs 3149

    3149 爱改名的小融 2  时间限制: 2 s  空间限制: 128000 KB  题目等级 : 黄金 Gold 题解       题目描述 Description Wikioi上有个人叫小融,他喜 ...

  4. POJ 3368.Frequent values-处理数据+RMQ(ST)

    昨天写的博客删了,占坑失败,还是先把RMQ玩的6一点再去搞后面的东西.废话少说,题解题姐姐_(:з」∠)_      Frequent values Time Limit: 2000MS   Memo ...

  5. PostgreSQL 二进制安装

    一.安装前准备工作 新建用户 sudo groupadd sql sudo useradd -g sql postgressudo passwd postgres 创建数据及日志目录,并做相应授权 s ...

  6. Codeforces 889C Maximum Element(DP + 计数)

    题目链接  Maximum Element 题意  现在有这一段求序列中最大值的程度片段: (假定序列是一个1-n的排列) int fast_max(int n, int a[]) { int ans ...

  7. 做IT这几年,我整理了这些干货想要送给你!

    没有一条路是容易的,特别是转行计算机这条路. 松哥接触过很多转行做开发的小伙伴,我了解到很多转行人的不容易,记得松哥大二时刚刚决定转行计算机,完全不知道这些东西到底应该怎么学,每天就是抱着书啃,书倒是 ...

  8. jQuery全选全不选

    $("[name='goods']").prop("checked", this.checked);

  9. jsonp _____跨域请求实现

    请求如下: $.ajax({ type: "GET", async:false, url: "http://127.0.0.1:8080/Cross-site-reque ...

  10. spring事务再次理解

    2.2.3 只读 事务的第三个特性是它是否为只读事务.如果事务只对后端的数据库进行该操作,数据库可以利用事务的只读特性来进行一些特定的优化.通过将事务设置为只读,你就可以给数据库一个机会,让它应用它认 ...