Intents, Entities, and Dialogs

[MUSIC] Welcome back we're now going to zero in on 0:00 the important things that you need to know to build a chat bot using 0:03 the conversation service that we created earlier. 0:08 >> The first thing we should do is get a bird's eye view of what we're building and 0:12 what the overall ecosystem Looks like. 0:18 We have provided you with a sample web app that will be the client in this. 0:21 The web app is using some Watson SDKs to make it very easy to communicate with 0:26 the Watson APIs available along the Watson Developer Cloud over here. 0:31 Users will input text into our simplified app, but 0:36 not that Watson can process text, voice, and visual input. 0:41 Our app will first pass the input data To a new conversation service that we create. 0:47 Each time we use a service from Watson, we create our own instance to connect to. 0:53 That allows us to train our unique instance with our own custom data. 0:59 The conversation service processes the input using its underlying 1:03 natural language processing technology. 1:08 Later we'll see how to communicate with more than just one service. 1:11 As we go along and build this part, 1:16 it will help to think about how it fits into this overall picture. 1:19 Then, now seems like a good time to point out that IBM has a variety of And 1:24 s available to make it easy for you to get started on a new p[roject. 1:30 No matter what you are working on, the are available 1:36 on and helpful are available on this page. 1:41 Both of these links are available for you in the teahers notes. 1:46 Back in our diagram, let's zoom into the conversation service. 1:51 Like a regular conversation, a conversation service with Watson has 1:56 some key elements in the speech or text that are needed for understanding. 2:01 The Watson conversation service defined these parts as intents, 2:07 entities, and dialog. 2:12 And they work together to deliver a conversational response to the user. 2:14 An intent represents the purpose of a user's input. 2:19 You define an intent for each type of user request you want your application to 2:23 support In the tool the name of an intent is always prefixed the pound character. 2:29 To train the workspace to recognize your intents you supply lots 2:35 of examples of user input and indicate which intents they left. 2:40 >> An entity represents a term or object that is relevant to your intents And 2:45 that provides a specific context for an intent. 2:51 For example, an entity might represent a city where the user wants to find 2:54 a business location or the amount of the bill payment. 2:58 In the tool the name of an entity is always prefix with the @ character. 3:01 To train the workspace to recognize your entities you list the possible values for 3:04 each entity and synonyms that users might enter. 3:09 >> A dialog is a brancing converstaion flow that defines 3:12 responses to the defined intents and entities. 3:17 You use the dialog building inthe tool to create 3:21 conversations with users to provide responses. 3:26