Chapter 10: Novel Trends on Graphs

In the previous chapters, we described different supervised and unsupervised algorithms that can be used in a wide range of problems concerning graph data structures. However, the scientific literature on graph machine learning is vast and constantly evolving and every month, new algorithms are published. In this chapter, we will provide a high-level description of some new techniques and applications concerning graph machine learning.

This chapter will be divided into two main parts – advanced algorithms and applications. The first part is mainly devoted to describing some interesting new techniques in the graph machine learning domain. You will learn about some data sampling and data augmentation techniques for graphs based on random walk and generative neural networks. Then, you will learn about topological data analysis, a relatively novel tool for analyzing high-dimensional data. In the second part, we will provide you with some interesting...

Technical requirements 

We will be using Python 3.6.9 for all our exercises. The following is the list of Python libraries that you must install for this chapter using pip. For example, you can run pip install networkx==2.5 on the command line, and so on:

networkx==2.5 
littleballoffur==2.1.8

All the code files relevant to this chapter are available at URL TO BE DECIDED.

Learning about data augmentation for graphs

In Chapter 8, Graph Analysis for Credit Card Transactions, we described how graph machine learning can be used to study and automatically detect fraudulent credit card transactions. While describing the use case, we faced two main obstacles:

• There were too many nodes in the original dataset to handle. As a consequence, the computational cost was too high to be computed. This is why we selected only 20% of the dataset.
• From the original dataset, we saw that less than 1% of the data had been labeled as fraudulent transactions, while the other 99% of the dataset contained genuine transactions. This is why, during the edge classification task, we randomly subsampled the dataset.

The techniques we used to solve these two obstacles, in general, are not optimal. For graph data, more complex and innovative techniques are needed to solve the task. Moreover, when datasets are highly unbalanced, as we mentioned in Chapter 8, Graph...

Learning about topological data analysis

Topological Data Analysis (TDA) is a rather novel technique that's used to extract features that quantify the shape of the data. The idea of this approach is that by observing how datapoints are organized in a certain space, we can reveal some important information about the process that generated it.

The main tool for applying TDA is persistent homology. The math behind this method is quite advanced, so let's introduce this concept through an example. Suppose you have a set of data points distributed on a space, and let's suppose you are "observing" them over time. Points are static (they do not move across the space); thus, you will observe those independent points forever. However, let's imagine we can create associations between these data points by connecting them together through some well-defined rules. In particular, let's imagine a sphere expanding from these points through time. Each point will...

Applying graph theory in new domains

In recent years, due to there being a more solid theoretical understanding of graph machine learning, as well as an increase in available storage space and computational power, we can identify a number of domains in which such learning theories are spreading. With a bit of imagination, you can start looking at the surrounding world as a set of "nodes" and "links." Our work or study place, the technological devices we use every day, and even our brain can be represented as networks. In this section, we will look at some examples of how graph theory (and graph machine learning) has been applied to, apparently, unrelated domains.

Graph machine learning and neuroscience

The study of the brain by means of graph theory is a prosperous and expanding field. Several ways of representing the brain as a network have been investigated, with the aim of understanding how different parts of the brain (nodes) are functionally or structurally...

Summary

In this chapter, we provided a high-level overview of some emerging graph machine learning algorithms and their applications for new domains. At the beginning of this chapter, we described, using the example provided in Chapter 8, Graph Analysis for Credit Card Transactions, some sampling and augmentation algorithms for graph data. We provided some Python libraries that can be used to deal with graph sampling and graph data augmentation tasks.

We continued by providing a general description of topological data analysis and how this technique has recently been used in different domains.

Finally, we provided several descriptions of new application domains, such as neuroscience chemistry, and biology. We also described how machine learning algorithms can also be used to solve other tasks, such as image classification, shape analysis, and recommendation systems.

This is it! In this book, we provided an overview of the most important graph machine learning techniques and...

Why subscribe?

• Spend less time learning and more time coding with practical eBooks and Videos from over 4,000 industry professionals
• Improve your learning with Skill Plans built especially for you
• Get a free eBook or video every month
• Fully searchable for easy access to vital information
• Copy and paste, print, and bookmark content

Did you know that Packt offers eBook versions of every book published, with PDF and ePub files available? You can upgrade to the eBook version at packt.com and as a print book customer, you are entitled to a discount on the eBook copy. Get in touch with us at [e-mail] for more details.

At www.packt.com, you can also read a collection of free technical articles, sign up for a range of free newsletters, and receive exclusive discounts and offers on Packt books and eBooks.