The 1st stage of my “Cognition Bot Taxonomy”.
The “No cognition” bot is purely mechanical and is not able to learn anything else than it is preprogrammed to do.
A “No cognition” chatbot would work as a switch:
You: asdkasf Bot: I did not understand you. Please write "on" or "off" You: on Bot: lamp is on You: off Bot: lamp is off
Python example (including source code): bot_1_no_congition.zip
Run the “bot_1_program.py” – requires Python 3.x
and type for example: “on”
“test_bot_1.py” is just for unit-testing.
A more complex algorithm would deal with more complex tasks, like a robot that cleans a room. For example, whenever the bot hits a wall, it turns 90° and continues:
The algorithm would have to be tested and improved, for it to deal with many types of rooms. For example, when it’s finished, it turns to a free path and starts over:
Two trips would have a good coverage:
The algorithm could be simplified, by adding randomness.
Whenever the bot hits a wall, it would turn in a random direction:
The bot will not learn anything new, but with randomness, it might solve problems we didn’t think off, and therefore seem like it is intelligent.