New hands-on guide demonstrates how to implement natural language processing business solutions
Natural Language Processing with AWS AI Services seeks to demystify NLP for just about anyone.
In Ali Baba and the Forty Thieves, Ali Baba overhears one of the thieves utter a magic phrase, “Open sesame,” which opens the mouth of a cave containing treasures.
Premkumar “Prem” Rangarajan, an artificial intelligence/machine learning specialist at Amazon Web Services (AWS), remembers his father reading this story to him as a child. “When I began working with artificial intelligence [AI] and natural language processing [NLP], this story came back to me,” he said. “I realized it was a fictional example of voice activation!”
Rangarajan says that today, AI/NLP can seem almost as magical as the secret code from the folktale.
Artificial intelligence is no longer an inaccessible technology. It’s no longer a career that requires us to study for 10 or 15 years of our lives and get multiple PhDs to begin.
“I mean, how do we even make computers, which can only understand ones and zeros, understand your voice?” he asks. “How does it understand that this sound means this with all of the tonal inflections, the accents, the languages? It was so fascinating, and that’s when it began for me. I’m fascinated with using voice for practical applications.”
In an effort to demystify some of that “magic”, Rangarajan and Mona Mona, an AI/ML specialist at Google and former AWS employee, wrote a book. Natural Language Processing with AWS AI Services is a hands-on guide which the authors say can help get any IT professional implementing AI/machine learning solutions before the monthly calendar flips to a new page.
“Artificial intelligence is no longer an inaccessible technology,” says Rangarajan. “It’s no longer a career that requires us to study for 10 or 15 years of our lives and get multiple PhDs to begin.
“Now you can actually understand and choose what you want and directly infuse these algorithms into your applications and build very powerful AI solutions. You don’t have to worry that if you have an idea today it will become a reality a year down the line. It can become a reality in one week, in two weeks.”
Spotting a need
Rangarajan says one area where artificial intelligence and machine learning can generate immediate business intelligence is in customer call centers.
“How can we improve the customer satisfaction scores? How can we understand whether the customer's issue was actually addressed in those conversations? How can we make the agents more efficient? How can we improve call closure rates?” he says. “We have the ability to use AI services to add intelligence to those conversations and to ensure that we address what the customer wants.”
Mona says soon after she began working at AWS, she recognized the power of the technology stack. “When I saw the power of these tools and I was introduced to some very interesting customer cases, I realized these services can provide natural language processing solutions quickly. You can build a chat bot, or an AI translation solution, or use NLP to do social media analytics. It’s all available to you.”
The authors had previously cowritten about a dozen AWS blog posts on AI/NLP, and from the comments they began to see a need for a new kind of book on NLP.
“We realized that a lot of books talk about the math and the science behind NLP, but there's not a lot of books that showed you how to apply the technology and actually solve the real-world need,” Rangarajan says.
A comprehensive lesson
The second section of the book demonstrates how NLP can be applied to business solutions, such as improving customer service, monetizing media content, extracting metadata from documents, and specific solutions for healthcare.
Finally, in section three, the book provides a hands-on guide to putting these solutions into production, including creating workflows and "building secure, reliable, and efficient NLP solutions”. In addition to the book, Rangarajan suggests anyone interested in AI/NLP to set up a free AWS tier account.
“The innovation needed to utilize artificial intelligence and natural language processing is already done by AWS with the AI services stack,” Rangarajan says. “You have AWS Comprehend, Amazon Translate, Amazon Transcribe, et cetera. All you have to do is make an API call to be able to access the intelligence behind those machine learning models.”
Mona notes that the book can be used in different ways by people with different roles within an organization.
“Suppose I'm a business executive. I don't want to read all the code. You could just read the appropriate business problem and solution chapter, the introduction, and the architecture proposed and summary. Then you can pass it on to your technical peers and say, ‘Now I see how it is done and I think this is what we need. Please go and build it,’” she says.
“On the other hand, if I'm a technical person, I will have a different perspective. I will literally read all the code. I can view the videos we have created for each code in the book. So, if you want to implement an end-to-end solution which your manager has given you, now you can go and implement it.”
A resource for career change
Rangarajan says the book is a good primer for someone wanting to transition to focus to AI/NLP, just as he did. He began as an IBM AS 400 programmer and then moved on to become an enterprise application integration architect. During that time he became interested in doing more with machine learning, which led to him joining AWS. Soon he developed an interest in NLP.
Around that time, Rangarajan was asked to work on a project for the celebration of the opening of a new AWS tech hub in Houston. He created an NLP project.
“There was something called ‘Simple Beer Service,’ and this provided an opportunity to upgrade it with Alexa. So, you say to Alexa, ‘Pour me a beer,’ and you use the password. Alexa will then control a Raspberry Pi device, open the beer lines, and pour the beer for you.”
That project (which drew the attention of Houston’s mayor) helped to cement his interest in pursuing NLP — and that interest eventually led to this book. Rangarajan said his own experiences helped shape his approach to the book.
“The book is good for students or working professionals who are interested in moving to an AI/machine learning career. That’s something that’s in demand, so it can be a profitable career move,” Rangarajan says.
The book, combined with the AWS Free Tier accounts, AWS Machine Learning University video tutorials, and, of course, the Amazon AI/NLP technology stack, can help ease entry into the field.
“Amazon's philosophy is that anyone can do this,” Mona says. “Even if you have no basic coding experience, you can still create a scalable application using these AI services. That is the goal, that any student or any IT professional can easily pick these services and implement infused, beautiful, innovative solutions and applications in a week's time. You don’t have to spend a lot of time learning.”
Looking ahead, Rangarajan is writing another book on cloud-native machine learning on AWS. “It’s going to be broad-scale AI and ML and cover the machine learning workflow. So, we are talking about algorithms, neural networks, and how different personnel use machine learning and AI within organizations.”
His mission to help others unlock the potential treasures of machine learning is certainly a goal of which Ali Baba would approve.