Home

Introduction

In this wiki, we will explain how to use the Ovation framework to build components of a chatbot. For Ovation Summer Academy (OSA) we have defined some tasks and business use-cases that the participating teams can choose from and work on during the course of the event. We have prepared these tasks such that they cover the scientific end of Conversational Intelligence (CI) and at the same time focus on realistic business scenarios that can make use of Conversational Intelligence.

We looked at the profiles of the participants and saw that we have a mix of people who have a good experience in CI and some who know about existing solutions but do not have a hands-on experience. We believe that it will be a huge learning experience for all the participants no matter what their experience level is. The following sections will explain how to prepare for OSA and what are the contents of this wiki.

How to prepare for OSA

To give you a heads up, and a buffer to prepare for OSA we recommend you to already look into the following:

1. Tensorflow: We use Tensorflow as the ML/DL backbone of the Ovation framework. We have integrated version 1.2.0 f Tensorflow in Ovation and recommend you to at least go through the following tutorials,

        a. Getting Started

        b. Getting Started With TensorFlow

        c. MNIST For ML Beginners

        d. Deep MNIST for Experts

2. Rasa: To implement business use-cases, we will use Rasa. It is an open-source CI framework which helps you developing chatbots faster and the best part, for free. We are looking forward to building some demos in OSA for the use-cases so we strongly recommend you to go through the following tutorial on integrating Rasa with Slack

        b. Training a Rasa bot: This will help you in developing the demos of the use-cases at OSA.

        b. Building a Slackbot with Rasa: This will help you in developing the demos of the use-cases at OSA.

Contents

The pages below describe what is present in the Ovation Framework. In PART I by describing what tasks we are focusing on, and follow with an example on how to use the framework. We then proceed (PART II) to a description of each of the Components of the Ovation Framework. Finally, in PART III we describe contribution guidelines.

PART I

1. Tasks: Tasks descriptions and how to address them;

2. Integrate With Rasa: Explains how you can integrate your models with Rasa.

3. Ovation by Example: An example of how to use the Ovation framework and its various modules;

PART II

4. Components of Ovation: Description of all the components of Ovation;

5. The Datasets API: Information on how to access data present in the supported datasets, and which datasets are available;

6. Build your own Model: How to create your own model using Ovation;

7. Using Templates: How to use existing templates to write your own model in minutes;

8. Ovation-voice: How to use the Ovation voice interface to build a voice based bot;

9. Tools: How to use some of the additional tools that Ovation provides;

PART III

10. Contribution Guidelines: How to contribute to the Ovation framework.

11. Pre Trained Models, Environment, Datasets and User Accounts: Some information about where the datasets are, what dependencies needs to be installed and what are the user accounts that the participants will used and what pre-trained models Ovation-CI provides.