A simple cell is just a pair of Input/Output, which tells the virtual robot how to respond to certain input. For more than 95% of projects, 99% of cells are simple cells. This means if you have learned simple cells, you're almost ready to build your virtual robots.
Example 1. A simple cell
- Input: Hello
- Output: Hi~~~
The Input field represents what users might input. The Output field describe what the brain should respond. Together they define a "behavior." For the above example cell, its behavior is to say "Hi~~~" when users say "Hello."
Why "Hi~~~"? We want to differentiate it from the default response provided by the system. When we get "Hi~~~," we know the cell above works.
To create the simple cell above,
- Create a brain if you haven't.
- Click Cells tab.
- Click icon.
- Enter "hello" for Input field.
- Click Submit button.
- Enter "Hi~~~" for Output field.
- Click Save button.
Note: When you create a simple cell, you may ignore all of the Advanced sections, which will be explored at a later stage.
The newly created cell is loaded to your virtual robot immediately. To verify if the cell works as expected,
- Click Training tab.
- Enter "Hello!"
- Click Submit or press Enter key.
The robot will return "Hi~~~" as you specified in the cell above.
- Hello!
- Hi~~~
There are a few things you need to know about Input field.
- Input represents what end users might say.
- It doesn't accept any punctuation points because we intentionally ignore points.
- It is case insensitive. Hello, HELLO or hello works the same for Input.
There are also a few key points you need know about Output.
- Output represents the robot’s reply to end users.
- It returns the text with punctuation points and cases as is.
Machine Learns
It's cool that you have "taught" your virtual robot to give a different response to "Hello" by creating a cell. But what if users say "Hi" instead of "Hello?" Still on the Training, you will get "Hi~~~" if you enter "Hi."
- Hi!
- Hi~~~
That's magical and the exact reason why you should use BrainShop for you AI project. You instruct the robot how to respond to "Hello" and then it is able to handle not only "Hello" but also "Hi", according to your instruction. There are a few key points behind such ability.
- The robot knows "Hello" and "Hi" largely mean the same thing.
- It learns from your cells by analogy.
So we have the following "golden" rules for you regarding working with simple cells.
- You need to create only one cell for each possible meaning that user might want to convey. You don't have to enumerate all possible sentences that convey the same meaning.