Bot 2 – Human cognition required – Supervised learning

The 2nd stage of my “Cognition Bot Taxonomy”.

The bot is able to create relations between sentences (but only exact matches), on observing human interaction:

You: turn on lamp

Robot is thinking:
> Known 'on' commands:  ['on']
> Known 'off' commands:  ['off']
> No match found

Robot: I did not understand you. Please write "on" or "off"

----

You: on

Robot is thinking:
> Known 'on' commands:  ['on']
> Known 'off' commands:  ['off']
> Match found.
> The last unknown command: 'turn on lamp', must mean 'on'!
> I have learned a new command: 'turn on lamp'

Robot: lamp is on

----

You: off

Robot is thinking:
> Known 'on' commands:  ['on', 'turn on lamp']
> Known 'off' commands:  ['off']
> Match found.
Robot: lamp is off

----

You: turn on lamp

Robot is thinking:
> Known 'on' commands:  ['on', 'turn on lamp']
> Known 'off' commands:  ['off']
> Match found.
Robot: lamp is on

----

You: off

Robot is thinking:
> Known 'on' commands:  ['on', 'turn on lamp']
> Known 'off' commands:  ['off']
> Match found.
Robot: lamp is off

----

You: turn on lamp, I can't see anything

Robot is thinking:
> Known 'on' commands:  ['on', 'turn on lamp']
> Known 'off' commands:  ['off']
> No match found

Robot: I did not understand you. Please write "on" or "off"

This type of bot is the kind of technology behind Web 2.0: Wikipedia, Facebook, YouTube and even Google Search Engine. They are built to harvest human input and human computing to create useful tools.


Python example (including source code): bot_2_human_congition_required.zip

Run the “bot_2_program.py” – requires Python 3.x
and type: “on”

“test_bot_2.py” is just for unit-testing.


View the full “Cognition Bot Taxonomy”

Add a Comment

Your email address will not be published.