Aflyst Computational creativity

- Please write me a poem about nature.

And the LLM returns a poem about nature.
 

 - Please write me a poem without using the letter E.

And the LLM returns a poem without using the letter E.
 

 - Please write me a poem in which the first letters of all words compose the word "kindness".

And the LLM returns a poem in which the first letters of all words compose the word "kindness".

We are in a time in which Large Language Models (LLM) achieve impressive results, especially in areas that seemed to be purely and innately human. As we can see, the computational model seems to resemble quite well the human ability of producing language and making it sound creative.

In this course, we want to take some steps back and try to understand what we mean by being creative. Once that theoretical grounds are established, we will go through a brief history of computational creativity models, starting from the simplest template-based ones to more complex ones. In this course we will discuss what can it mean to have a good computational creativity model by discussing how it should be evaluated. Is it enough to resemble human creativity? Does it need to be good? What is it to be good? What about novelty?