Reader small image

You're reading from  Conversational AI with Rasa

Product typeBook
Published inOct 2021
PublisherPackt
ISBN-139781801077057
Edition1st Edition
Tools
Right arrow
Authors (2):
Xiaoquan Kong
Xiaoquan Kong
author image
Xiaoquan Kong

Xiaoquan is a machine learning expert specializing in NLP applications. He has extensive experience in leading teams to build NLP platforms in several Fortune Global 500 companies. He is a Google developer expert in Machine Learning and has been actively involved in contributions to TensorFlow for many years. He also has actively contributed to the development of the Rasa framework since the early stage and became a Rasa Superhero in 2018. He manages the Rasa Chinese community and has also participated in the Chinese localization of TensorFlow documents as a technical reviewer.
Read more about Xiaoquan Kong

Guan Wang
Guan Wang
author image
Guan Wang

Guan is currently working on Al applications and research for the insurance industry. Prior to that, he was a machine learning researcher at several industry Al labs. He was raised and educated in Mainland China, lived in Hong Kong for 10 years before relocating to Singapore in 2020. Guan holds BSc degrees in Physics and Computer Science from Peking University, and an MPhil degree in Physics from HKUST. Guan is an active tech blogger and community contributor to open source projects including Rasa, receiving more than10,000 stars for his own projects on Github.
Read more about Guan Wang

View More author details
Right arrow

Chapter 10: Conversation-Driven Development and Interactive Learning

Compared with traditional software development, the challenge of developing a chatbot is far greater. This is largely due to the fact that the user could say anything to the dialogue bot. Of course, as a developer, you cannot cope with all possible situations for your robot. Therefore, it is extremely important to understand your user's queries.

In this chapter, we will introduce a methodology in which to develop a dialogue system called Conversation-Driven Development (CDD). This methodology improves dialogue robots by observing, summarizing, and modifying the dialogue process. Additionally, we will introduce a tool for CDD: Rasa X. In a step-by-step manner, we will learn how to use Rasa X to complete all stages of CDD. Finally, we will also introduce you to Interactive Learning, which is a technical solution that allows developers to interact with the dialogue system to test system capabilities and quickly...

Introduction to CDD

CDD is a methodology that enables you to develop a dialogue system; it was introduced by the Rasa team. It is an iterative and interactive process: developers observe the behavior of users and improve chatbot performance based on those observations.

CDD involves the following steps:

  1. Sharing your bot: We should distribute our product prototype for user testing as soon as possible. No matter how hard developers try, users will always have something new to input into the chatbot. Many teams spend months developing chatbots and focusing on conversations that, in reality, users never have.
  2. Reviewing conversations: We should spend time studying the conversation between users and our chatbot. It is very helpful to study real user conversations at each stage of development (from the prototype to the real product). Far too many teams only focus on simple attributes, such as how many users express certain intentions and so on. Instead, they should spend more...

Introduction to Rasa X

Rasa X is a toolset for CDD and was developed by the Rasa team.

The license terms of Rasa X

Rasa X, as mentioned in this book, refers to Rasa X Community Edition. Rasa X is a free, closed source tool that is available to all developers. The use of Rasa X requires you to accept its license terms: https://storage.googleapis.com/rasa-x-releases/rasa_x_ce_license_agreement.pdf. Rasa X Community Edition is free for non-commercial use. It is also free for commercial use, as long as you don't provide it as a service (Software as a Service) to others. For more details please refer to the official license.

Installing Rasa X

Rasa X is a tool for production environments, so the official documentation (https://rasa.com/docs/rasa-x/installation-and-setup/installation-guide) offers many installation methods (such as local installation, Helm Chart installation, and Docker Compose installation). In this chapter, we will only introduce one of the installation...

Summary

In this chapter, we discussed CDD, which is a methodology that is used to construct dialogue systems efficiently. We introduced a tool for CDD: Rasa X. We explained, in detail, how to use Rasa X to complete the six steps of CDD, that is, sharing, reviewing, annotating, testing, tracking, and fixing. Additionally, we discussed interactive learning and demonstrated, in detail, how to use the Rasa CLI to complete interactive learning. After studying these two topics, you should now have more confidence regarding how to build a successful dialogue system in theory and in practice.

In the next chapter, we will discuss how to debug and optimize the dialogue system and introduce Rasa's open source community ecosystem.

lock icon
The rest of the chapter is locked
You have been reading a chapter from
Conversational AI with Rasa
Published in: Oct 2021Publisher: PacktISBN-13: 9781801077057
Register for a free Packt account to unlock a world of extra content!
A free Packt account unlocks extra newsletters, articles, discounted offers, and much more. Start advancing your knowledge today.
undefined
Unlock this book and the full library FREE for 7 days
Get unlimited access to 7000+ expert-authored eBooks and videos courses covering every tech area you can think of
Renews at $15.99/month. Cancel anytime

Authors (2)

author image
Xiaoquan Kong

Xiaoquan is a machine learning expert specializing in NLP applications. He has extensive experience in leading teams to build NLP platforms in several Fortune Global 500 companies. He is a Google developer expert in Machine Learning and has been actively involved in contributions to TensorFlow for many years. He also has actively contributed to the development of the Rasa framework since the early stage and became a Rasa Superhero in 2018. He manages the Rasa Chinese community and has also participated in the Chinese localization of TensorFlow documents as a technical reviewer.
Read more about Xiaoquan Kong

author image
Guan Wang

Guan is currently working on Al applications and research for the insurance industry. Prior to that, he was a machine learning researcher at several industry Al labs. He was raised and educated in Mainland China, lived in Hong Kong for 10 years before relocating to Singapore in 2020. Guan holds BSc degrees in Physics and Computer Science from Peking University, and an MPhil degree in Physics from HKUST. Guan is an active tech blogger and community contributor to open source projects including Rasa, receiving more than10,000 stars for his own projects on Github.
Read more about Guan Wang