A summary of SVMs - A Practical Consequence of Learning Theory by Bernhard Scholkopf
Nicholas M. Synovic
- 4 minutes read - 670 wordsA summary of SVMs: A Practical Consequence of Learning Theory
Bernhard Scholkopf; DOI
For the summary of the paper, go to the Summary section of this article.
Table of Contents
First Pass
Discussion about the title, abstract, introduction, section and sub-section headings, and conclusion
The essay SVMs - a practical consequence of learning theory by Bernhard Scholkopf (as part of the larger Support vector machine collection of essays in the July/ August edition of the 1998 IEEE Intelligent Systems magazine) [1] discusses the underlying theory that powers Support Vector Machine (SVM) algorithms and argues that these algorithms are useful and performant. His essay contains sections on Learning pattern recognition from examples, Hyperplanes, Feature spaces and kernels, SVMs, and Current developments and open issues which indicates an essay that will holistically look at SVMs, rather than a particular facet of them.
Category
What type of paper is this work?
This paper is best classified as an informative essay on the benefits of SVMs from a theoretical and practical view.
Context
What other types of papers is the work related to?
This paper is most related to other papers within the magazine’s collection, as well as work that goes into the theory behind SVMs.
Contributions
What are the author’s main contributions?
A brief description of the theory that powers SVMs, as well as identifying where SVMs are practical.
Second Pass
Background Work
What has been done prior to this paper?
Work has been done to develop and implement the SVM algorithm.
Motivation
Why should we care about this paper?
Because it provides a concise description of the theory that powers SVMs, and practical usages of SVMs.
Figures, Diagrams, Illustrations, and Graphs
Are the axes properly labeled? Are results shown with error bars, so that conclusions are statistically significant?
The paper doesn’t provide and graphs or charts. However, the figures and diagrams that are presented are clearly explained in the descriptions, are well made, and are easy to comprehend.
Clarity
Is the paper well written?
Sort of? The theory components of the essay are written distinctly differently than the introduction and concluding sections of the paper. This could be due to the discussion of mathematical prose; but due to this, the essay has two different voices.
Relevant Work
Mark relevant work for review
The following relevant work can be found in the Citations section of this article.
- The Nature of Statistical Learning Theory [2]
Future Directions
My own proposed future directions for the work
I’d like to explore the usage of SVMs for face or object detection and compare it against the usage of DL techniques on both traditional and low-powered metrics.
Summary
A summary of the paper
The essay SVMs: A Practical Consequence of Learning Theory by Bernhard Scholkopf (as part of the larger Support vector machine collection of essays in the July/ August edition of the 1998 IEEE Intelligent Systems magazine) [1] discuss both the mathematical theory and current practice of using SVMs. SVMs are useful in a research aspect as their functionality can be mathematically explained. SVMs are a linear classifier that operate in multi-dimensional space through the usage of a hyper plane. Hyper planes are chosen by finding support vectors, which are instances of a class that are closest to one another. The hyper plane then splits these two instances into two separable sides. To assist in this calculation, a kernel algorithm is applied to map one multi-dimensional space to another for easier computation.
Overall, this paper provides a good understanding of the theory behind SVMs. It also alludes to additional usages of SVMs and their current problems, but it is not focused on discussing or resolving them.
Summarization Technique
This paper was summarized using a modified technique proposed by S. Keshav in his work How to Read a Paper [0].