Why Deep Learning Works – Key Insights and Saddle Points
Why Deep Learning Works – Key Insights and Saddle Points
A quality discussion on the theoretical motivations for deep learning, including distributed representation, deep architecture, and the easily escapable saddle point.
This post summarizes the key points of a recent blog post by Rinu Boney, based on a lecture by Dr. Yoshua Bengio from this year's Deep Learning Summer School in Montreal, which discusses the theoretical motivations for deep learning.

"To generalize locally, we need representative examples for all relevant variations."
Deep learning is about learning multiple levels of representations, corresponding to multiple levels of abstractions. If we are able to learn these multiple levels of representation, we are able to generalize well.
After setting the general tone of the post with the above (paraphrased) statement, the author presents a number of different artificial intelligence (AI) strategies, from rule-based systems to deep learning, and notes on at which levels of learning their components function. He then states the 3 keys to moving from machine learning (ML) to true artificial intelligence: Lots of data, very flexible models, and powerful priors, and that, since classical ML can handle the first 2, his post deals with the third.
On the path toward AI from today's ML systems, we need learning, generalization, ways to fight the curse of dimensionality, and the ability to disentangle the underlying explanatory factors. Before explaining why non-parametric learning algorithms won't get us to true AI, he gives a nuanced definition of non-parametric. He explains why smoothness, a classical non-parametric approach, won't work on high-dimensionality, and then provides the following insight re: dimensionality:
"If we dig deeper mathematically, it's not the number of dimensions but the number of variations of functions that we learn. In this case, smoothness is about how many ups and downs are present in the curve."

"A line is very smooth. A curve with some ups and downs is less smooth but still smooth."
So, it's clear that smoothness will not beat the curse of dimensionality alone. In fact, smoothness doesn't even apply to modern, complex problems like computer vision or natural language processing. After discussing the downfalls of such competing methods as Gaussian kernels, Boney sets his sights on moving past smoothness, and why that's necessary:
"We want to be non-parametric in the sense that we want the family of functions to grow in flexibility as we get more data. In neural networks, we change the number of hidden units depending on the amount of data."
He notes that in deep learning, 2 priors are used, namely distributed representations and deep architecture.
Why distributed representations?
"With distributed representations, it is possible to represent exponential number of regions with a linear number of parameters. The magic of distributed representation is that it can learn a very complicated function (with many ups and downs) with a low number of examples."
In distributed representations, features are individually and independently meaningful, and they remain so regardless of what the other features are. There maybe some interactions but most features are learned independent of each other. Boney states that neural networks are very good at learning representations capturing the semantic aspects, and that their generalization power is derived from these representations. As a practical exploration of the topic, he recommend's Cristopher Olah's article for some information on distributed representation and Natural Language Processing.
There is a lot of misunderstanding about what depth means.
"Deeper networks does not correspond to a higher capacity. Deeper doesn't mean we can represent more functions. If the function we are trying to learn has a particular characteristic obtained through composition of many operations, then it is much better to approximate these functions with a deep neural network."
Boney then comes full circle. He explains that one of the reasons neural network research was abandon (once again) in the late 90s was because the optimization problem is non-convex. The realization from the work in the 80s and 90s that neural networks have an exponential number of local minima, along with the breakout success of kernel machines, also led to this downfall, as did the fact that networks may get stuck on poor solutions. Recently we have evidence that the issue of non-convexity may be a non-issue, which changes its relationship vis-a-vis neural networks.
"A saddle point is illustrated in the image above. In a global or local minima, all the directions are going up and in a global or local maxima, all the directions are going down."
Saddle Points.
"Let us consider the optimization problem in low dimensions vs high dimensions. In low dimensions, it is true that there exists lots of local minima. However in high dimensions, local minima are not really the critical points that are the most prevalent in points of interest. When we optimize neural networks or any high dimensional function, for most of the trajectory we optimize, the critical points(the points where the derivative is zero or close to zero) are saddle points. Saddle points, unlike local minima, are easily escapable."
The intuition with the saddle point, is that, for a minima located close to the global minima, all directions should be climbing upward; going further downward is not possible. Local minima exist, but are very close to global minima in terms of objective functions, and theoretical results suggest that some large functions have their probability concentrated between the index (the critical points) and the objective function. The index is the fraction of directions moving downward; for all values of index not 0 or 1 (local minima and maxima, respectively), then it is a saddle point.
Boney goes on to say that there has been empirical validation corroborating this relationship between index and objective function, and that, while there is no proof the results apply to neural network optimization, some evidence suggests that the observed behavior may well correspond to the theoretical results. Stochastic gradient descent, in practice, almost always escapes from surfaces that are not local minima.
This all suggests that local minima may not, in fact, be an issue because of saddle points.
Boney follows his saddle points discussion up by pointing out a few other priors that work with deep distributed representations; human learning, semi-supervised learning, and multi-task learning. He then lists a few related papers on saddle points.
Rinu Boney has written a detailed piece on the motivations for deep learning, including a good discussion on saddle points, all of which is difficult to do justice with a few quotes and some summarization. If you are interested in a deeper discussion of the above points to visit Boney's blog and read the insightful and well-written piece yourself.
Bio: Matthew Mayo is a computer science graduate student currently working on his thesis parallelizing machine learning algorithms. He is also a student of data mining, a data enthusiast, and an aspiring machine learning scientist.
Related:
Most popular last 30 days
Most viewed last 30 days
- Top 5 arXiv Deep Learning Papers, Explained - Oct 1, 2015.
- Data Lake vs Data Warehouse: Key Differences - Sep 29, 2015.
- 60+ Free Books on Big Data, Data Science, Data Mining, Machine Learning, Python, R, and more - Sep 4, 2015.
- R vs Python for Data Science: The Winner is ... - May 26, 2015.
- R vs Python: head to head data analysis - Oct 13, 2015.
- 30 Cant miss Harvard Business Review articles on Data Science, Big Data and Analytics - Sep 30, 2015.
- 9 Must-Have Skills You Need to Become a Data Scientist - Nov 22, 2014.
Most shared last 30 days
- Top 5 arXiv Deep Learning Papers, Explained - Oct 1, 2015.
- 30 Cant miss Harvard Business Review articles on Data Science, Big Data and Analytics - Sep 30, 2015.
- 90+ Active Blogs on Analytics, Big Data, Data Mining, Data Science, Machine Learning - Oct 8, 2015.
- How Big Data Helps Build Smart Cities - Oct 16, 2015.
- Does Deep Learning Come from the Devil? - Oct 9, 2015.
- 5 steps to actually learn data science - Oct 6, 2015.
- R vs Python: head to head data analysis - Oct 13, 2015.
Why Deep Learning Works – Key Insights and Saddle Points的更多相关文章
- why deep learning works
https://medium.com/towards-data-science/deep-learning-for-object-detection-a-comprehensive-review-73 ...
- Growing Pains for Deep Learning
Growing Pains for Deep Learning Advances in theory and computer hardware have allowed neural network ...
- Decision Boundaries for Deep Learning and other Machine Learning classifiers
Decision Boundaries for Deep Learning and other Machine Learning classifiers H2O, one of the leading ...
- Why deep learning?
1. 深度学习中网络越深越好么? 理论上说是这样的,因为网络越深,参数也越多,拟合能力也越强(但实际情况是,网络很深的时候,不容易训练,使得表现能力可能并不好). 2. 那么,不同什么深度的网络,在参 ...
- Use of Deep Learning in Modern Recommendation System: A Summary of Recent Works(笔记)
注意:论文中,很多的地方出现baseline,可以理解为参照物的意思,但是在论文中,我们还是直接将它称之为基线,也 就是对照物,参照物. 这片论文中,作者没有去做实际的实验,但是却做了一件很有意义的事 ...
- (转)WHY DEEP LEARNING IS SUDDENLY CHANGING YOUR LIFE
Main Menu Fortune.com E-mail Tweet Facebook Linkedin Share icons By Roger Parloff Illustration ...
- The Brain vs Deep Learning Part I: Computational Complexity — Or Why the Singularity Is Nowhere Near
The Brain vs Deep Learning Part I: Computational Complexity — Or Why the Singularity Is Nowhere Near ...
- What are some good books/papers for learning deep learning?
What's the most effective way to get started with deep learning? 29 Answers Yoshua Bengio, ...
- (转) Learning Deep Learning with Keras
Learning Deep Learning with Keras Piotr Migdał - blog Projects Articles Publications Resume About Ph ...
随机推荐
- 推荐一款开源的C#TCP通讯框架
原来收费的TCP通讯框架开源了,这是一款国外的开源TCP通信框架,使用了一段时间,感觉不错,介绍给大家 框架名称是networkcomms 作者开发了5年多,目前已经停止开发,对于中小型的应用场景,够 ...
- phpmailer 参数使用说明
$AltBody--属性 出自:PHPMailer::$AltBody 文件:class.phpmailer.php 说明:该属性的设置是在邮件正文不支持HTML的备用显示 AddAddress--方 ...
- [原]iBatis.Net(C#)系列一:简介及运行环境
转载请注明http://www.cnblogs.com/13590/archive/2013/02/27/2934580.html 摘要:介绍iBatis.Net的基本情况和运行原理,运行环境中各参数 ...
- GEOS库学习之四:几何关系判断
原理上一篇已经介绍过了,这篇就直接进行程序练习 #include "geos.h" GeometryFactory factory; //创建一条环线,与线的区别就是环线是闭合的. ...
- 在matlab中将处理结果输出为shp文件
在matlab中读入shp文件很简单,一个函数shaperead就可以了,但输出为shp文件就稍微麻烦一些了.shp文件实际上就是一个struct,因此得到处理结果后,要先将数据变成struct结构, ...
- C语言操作符优先级
C语言操作符优先级 优先级 运算符 含 义 要求运算 对象的个数 结合方向 1 () [] -> . 圆括号 下标运算符 指向结构体成员运算符 结构体成员运算符 自左至右 2 ! 逻辑非运 ...
- Java 环境下使用 AES 加密的特殊问题处理
在 Java 环境下使用 AES 加密,在密钥长度和字节填充方面有一些比较特殊的处理. 1. 密钥长度问题 默认 Java 中仅支持 128 位密钥,当使用 256 位密钥的时候,会报告密钥长度错误 ...
- [CareerCup] 3.5 Implement Queue using Two Stacks 使用两个栈来实现队列
3.5 Implement a MyQueue class which implements a queue using two stacks. LeetCode上的原题,请参见我之前的博客Imple ...
- Linux 读书笔记 二
一.实验说明 1. 环境登录 无需密码自动登录,系统用户名shiyanlou,密码shiyanlou 若不小心登出后,直接刷新页面即可 2. 环境使用 完成实验后可以点击桌面上方的“实验截图”保存并分 ...
- 客户端禁用cookies后session是否还起效果
设置session和cookies的代码(webform1.aspx) if (txtName.Text == "wlzcool") { Session["uid&quo ...
Previous post