We’re very delighted to talk with Dr. Alessandro Negro, the Chief Scientist of GraphAware, who authored the recently published book, Graph-Powered Machine Learning.
Dr. Negro has been a long-time member of the graph community, and was the main author of the very first recommendation engine based on Neo4j. At GraphAware, he specializes in natural language processing (NLP), recommendation engines and graph-aided search.
During this Q&A, Alessandro shared some of the insights from his new book, and what readers should expect and gain from using it as a resource.
Amy Hodler: Why did you write the book? What motivated you to get started on it?
Dr. Alessandro Negro: Good question. Actually, I wrote this book, first of all for myself. I needed a mechanism for re-organizing my ideas, my thoughts and my experience in these amazing areas – that is, the conjunction between graphs and machine learning. And specifically inside GraphAware, we wanted a mechanism also to help newcomers quickly learn how to deal with the complex tasks related to graphs and machine learning.
On the other side, I needed also a mechanism for proving my ideas and the book provided me with the motivation for testing a few ideas I had, or that I actually experienced with some customers we’ve served over the years.
Hodler: So graphs and machine learning – big concepts. How do you see the role of graphs in the machine learning space?
Dr. Negro: Let me say that the role of graphs changed quickly in the last few years, and I think that it will change quickly even in the future.
So many years ago, when I started working with graphs and Neo4j, you know, there were a lot of people just curious about a new mechanism for getting to the classical relationships under the hood. So at the beginning, there were all of these possible scenarios and people just wanted to test this approach.
Over the years, things have changed a lot. Now the world is totally different, and people are experts in graphs. They now use graphs for more advanced services or more advanced tasks. And specifically, in this sense, machine learning is one of the most relevant because they needed a mechanism for accessing data and for analyzing data that was different than before.
Graphs are playing a key role, because graphs help to organize your data and even your predictive models. You can leverage some algorithms that are graph-specific – and this definitely is opening up new opportunities and bringing in new ideas.
So I’m expecting that we’ll see a growing number of people using graphs for solving their machine learning problem in the future, and it will be even more constant as an approach for solving even more complex problems.
Hodler: So your book is available through an early access program now from Manning the publisher, but also an excerpt from Neo4j. We know people are downloading it and seeing it. What kind of early feedback are you getting?
Dr. Negro: People are appreciating the topic. First of all, because more than one piece of feedback I received was about the fact that this was the missing book – a conjunction ring between two worlds that clearly work very well together.
There is a lot of literature on this topic, but none are structured in a way in which people can read about an idea, but also practice this idea. So that’s definitely the most interesting feedback I’ve gotten, and I’m constantly receiving feedback from people contacting me on LinkedIn and other channels.
Hodler: Is there anything else that you want to share with the audience that we haven’t discussed yet?
Dr. Negro: Well, I really hope that people will not read my book cover to cover, but actually use the book. You can definitely take what you need and use it in a way you prefer. The topics are not only theoretical topics, they are concrete use cases. They are also concrete datasets that you can use as an example for practicing ideas and applying them to a concrete problem.
That is the most important part: This book is definitely meant to be used – not just read through and posed on your desk.
Hodler: Well, you guys heard it here. A book that is meant to be used: Graph-Powered Machine Learning by Dr. Alessandro Negro.
Readers, you can get your free excerpt by clicking the link below. So thank you again, Alessandro, really appreciate it, and thank you for writing the book.
Dr. Negro: Thank you, Amy. Thank you, Neo4j.
