- Course Overview 4:07
- What is Watson Used For? 3:13
- Trying it Out 5:10
- IBM Cloud and a Sample Chatbot 5:00
- How Conversation and Discovery Services Work Together 4:19
- How Conversation and Tone Analyzer Services Work Together 3:33
- Starting Our Project on IBM Cloud 7:12
- Starting Our Local Client 6:28
- Review: What is Watson? 6 questions
Preview
Video Player
00:00
00:00
Starting Our Project on IBM Cloud
7:12Let's get started on IBM Cloud by setting up a new Conversation Workspace.
Like we said earlier,
we're gonna build this chatbot in Node.js.
0:00
We definitely invite you to follow along
with your programming language though.
0:03
As you'll see the Blue Mix
documentation is great and
0:07
contains examples in Node, Java,
and Python, which makes it easy for
0:10
you to do the same thing we're
doing in a different language.
0:13
Now Armen and I are working on local
Node installations using a text editor.
0:17
It's very easy to set up, so
if you need to do that check out
0:21
the installation instructions we
included in the teacher's notes.
0:24
It will walk you through setting up
Node and NPM on a Mac, Windows or Linux.
0:27
And if you need a text editor,
I'd recommend Atom from GitHub,
0:31
which is free and easy to use,
and also linked below.
0:34
All right, so once you have Node
set up on your local computer,
0:36
you can download the project starter
files from the teacher's notes and
0:39
use them to follow along.
0:42
Now if you prefer to use
Treehouse workspaces for this,
0:44
you can launch one using
the button on this page.
0:47
It will load a workspace with the same
project files we're going to use locally,
0:49
so you should be able to
follow along directly.
0:52
The only issue is that currently we
have a bug in how .env files are edited.
0:55
In workspaces you need to add .txt to
the file name to make some changes and
1:00
then you have to remove that
.txt extension and Save it and
1:04
then you can use it in the app itself.
1:07
We are essentially working through
the steps outlined in this training guide
1:10
in the Watson Docs.
1:13
This link is in the teacher's notes and
it would be helpful for
1:15
you to bookmark it to refer to later for
review.
1:18
I've just created a brand new account,
and I've confirmed it through email.
1:21
And now we're at this page where
we need to create an organization.
1:25
So Armen,
can you help me set this up correctly?
1:27
>> Gladly Ben.
1:30
A few steps that we want
to pay attention to.
1:31
Notice the region that is selected for
you happens to be US South.
1:35
We are going to keep that,
that is the region we are closest to.
1:39
And we need to create an organization.
1:42
If below you can see there
are two suggested names.
1:45
One of them is your email, and the other
one is your shorter form of email.
1:49
I recommend picking the ones just right
there as opposed to typing a new one,
1:52
whichever you'd like.
1:56
>> Sounds good.
1:57
>> Now we're going to create a space,
2:00
again, notice the suggestions down there,
development, test, production.
2:02
Pick one of those.
2:06
>> Lets do dev.
2:08
Yeah, that makes sense.
2:08
Okay, it looks like we
are set up with the basics.
2:10
Click here?
>> Very well, yes.
2:14
>> And we are coming up on the dashboard.
2:16
All right mine does not
have as much as yours did.
2:18
>> A good place to go typically at this
time, Ben, is to click the Catalog tab.
2:22
And now, we have a great view of all
the services that are on our platform.
2:29
>> Cool, yeah, like you mentioned before,
more than just Watson,
2:35
there are a lot of other
services available.
2:37
But I'm guessing I want to click
Watson down here at the bottom of
2:40
the platform site.
2:42
>> For our purposes, let's go to
the Watson set of cognitive services.
2:43
And let's just take a look at
the 12 services that are here.
2:48
We'll be working with
the Conversation service,
2:53
we'll also work with
the Discovery service, and
2:55
another one of our favorites
is the Tone Analyzer.
2:59
Interesting to note that
the Natural Language Classifier, Ben,
3:05
is included in the Conversation service.
3:10
>> Right right.
3:12
>> That really helps in disambiguating and
3:13
taking care of the nuances of our speech,
our colloquial way of speaking.
3:17
>> So that's the magic behind the scenes.
3:22
We're using the Conversation service,
which is a wrapper for
3:23
all of the other work that
Watson is doing for us.
3:26
>> Absolutely.
3:28
Another good point to note is that NLC,
3:29
that's your language classifier,
is trained on Wikipedia.
3:33
So you're already getting
machine learned set of
3:40
training data where,
from there on, we can go on.
3:45
And it allows us to speak comfortably,
colloquially,
3:50
and it would understand
our nuances in our speech.
3:54
>> Very cool.
3:57
All right so,
I should click on Conversation, right?
3:59
>> Yes,
let's start with the Conversation service.
4:00
Wonderful, a good habit is
to give the service a name,
4:05
typically I leave Conversation intact and
just backspace over the 95 and
4:09
something meaningful to you,
your name or whatnot.
4:14
This is our Ben and Armen together.
4:18
Here we go.
4:22
>> Wonderful, and really,
that's all we need.
4:22
Let's click Create.
4:24
>> I see it's got organization and
the dev space we created.
4:26
>> Absolutely.
4:29
>> Let it spin for a moment.
4:30
All right, looks like we have it.
4:34
>> Wonderful Ben, let's go ahead and
4:36
a few things I'd like to show you
before we do click on the Launch tools.
4:39
Notice the Service
credentials link to the left.
4:42
Let's explore that.
4:47
Further down, you're seeing View
credentials to the bottom of the page.
4:50
>> This is gonna important shortly.
4:54
>> You're absolutely right,
that username and
4:56
password is really our currency as
we're dealing with RESTful services.
4:59
We will be copying pasting
that in a number of files.
5:03
Very well back to the Manage,
and let's click Launch tool.
5:08
>> Okay so it's asking me to login again,
but hopeful it retains my credentials and
5:15
lets me through.
5:20
>> And it's true, yes.
5:21
Interesting to note that out
5:23
of the box you are presented
with a Car Dashboard sample app.
5:26
>> Very helpful.
5:31
>> Very helpful, it allows us to get
started with an already built Conversation
5:32
full of entities, intents, and so forth.
5:37
[COUGH] In early example we're
going to create a new workspace.
5:40
>> Cool, and we'll come back to that
sample later because we do wanna look at
5:45
some of the stuff in there.
5:48
But for the new one,
what kind of name should I give this?
5:49
I guess this is a different car dashboard?
5:52
>> Correct.
>> This is the Treehouse Car Chatbot.
5:55
And I can give it an optional
description if I want,
5:57
working on the course with Armen,
and Create.
6:02
>> Notice Ben,
you are presented by a workspace.
6:07
A workspace is a container full
of the artifacts that define
6:11
the conversation flow for an application.
6:15
The natural language processing for
6:19
the conversation service happens inside a
workspace and you should add information.
6:21
The workspace trains itself,
as we will note,
6:27
the automatic training of the system
that takes place in the background.
6:31
You don't have to do
anything to initiate it.
6:35
This will make more sense as
we add different parts to it.
6:37
We should also note that
a single Conversation
6:42
service instance can contain
multiple workspaces.
6:44
Here is an example of why you may
want to have more than one workspace.
6:49
Perhaps one of your workspaces is for
chit-chat.
6:54
The greetings and the hello's and
goodbye's,you can keep that separate from
6:58
the more germane and core topic of
your conversation in a separate.
7:02
>> That's a good idea, I love well
organized components like that, very cool.
7:07