I'd like to know more about implementing emotional intelligence.

Given I'm implementing a chatbot and I'd like to introduce the levels of curiosity to measure whether user text input is interesting or not.

A high level would mean the bot is asking more questions and is following the topic. A lower level of curiosity makes the bot not asking any questions and changing the topics.

Less interesting content could mean the bot doesn't see any opportunity to learn something new or it doesn't understand the topic or doesn't want to talk about it, because of its low quality.

How this possibly can be achieved? Are there any examples?


2 Answers 2


It's possible to implement a form of curiosity-driven behavior without requiring full 'emotional intelligence'. One elementary strategy would be to define some form of similarity measure on inputs.

More generally, Jurgen Schmidhuber has pioneered work on 'Artificial Curiosity/Creativity' and 'Intrinsic Motivation' and has written a number of papers on the subject:

Here is a video of a nice associated presentation.


I think "curiosity" in AI would signify a 'desire to search.' It's an interest, that is experienced by some agent, in making something known that was previously unknown.

So to define how much curiosity a chat bot should have, we should:

  1. Specify what kinds of information the agent prefers knowing.
  2. Measure how much information is unknown about those preferred subjects. ('what is the user's name?' or 'What does the user need help with?')
  3. Measure the difficulty in making each unknown fact known.
  4. Sort unknown facts by difficulty of finding the answer.
  5. Set the "desire to search" on the highest ranking unknown fact.

While simplistic, those steps would constitute a state of affairs sufficient to describe "curiosity," in my opinion.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .