DMIND stands for “Digitalized Mind”. The best thing about it is that it is not about some random mind out there, but, for example, Shakespeare’s mind – the greatest writer in English literature. The samples that you can see below are not short stories and poems written by a human, nor they were written by Shakespeare himself as much as they remind of his writing style.
They are written by DMIND, a digitalized, “Shakespearized”, fine-tuned silicon-based creature that uses artificial intelligence that has adopted Shakespeare’s writing style. The samples on the pages listed below were created when this artificial neural network has transferred Shakespeare’s voice into topics as contemporary as the topic about COVID 19 is and old as time as the topic of Christmas.
In other words, the stories and poems written about COVID 19 and Christmas that you can see below, were written by an artificial neural network reflecting Shakespeare’s narrative voice.
Example of Autogenerated Poetry in Shakespeare's Style
O where thou livest, who art thou,
And where thou art dead?
What dost thou see in thee, what dost thou make?
If thou be an inhabitant,
If thou have no eyes to see,
What dost thou see? what is this?
What is the matter, wherewith thou art?
Which is thy life’s work?
Or what canst thou say,
For what work, thou art to be?
To think, thou art to act,
To do what?
What dost thou say, what art thou to do?
Thou art in the sight of death,
Thou art in the sight of life,
Who canst thou see, who canst thou see?
When thou dost live, thou shalt never rest.
A Little Bit About the Technology Behind
So now you may be asking yourself, what is the exact logic behind style transfer and how it is possible to write contemporary stories in the style of Shakespeare? Well, regardless of whether it may sound like some kind of magic, it is actually super easy to understand. In fact, this technology is already seen in image processing, making certain images look as if they were made by famous painters such as Van Gogh himself.
To be more precise in a technical sense, what truly style transfer represents are the internal capabilities of neural networks, pre-trained with the creations of famous artists. When it comes to training neural networks, they serve to determine two distance functions. The first one describes the differences in the content of the two texts, and the other one determines the differences between the styles of those two texts. After those two distance functions are determined, neural networks use backpropagation to minimize the differences between them. When the process is finished, you then have a mixture of the given content and the given style transferred into one single text.
Examples of Writings for Topics Shakespeare Never Wrote About
The best thing is that despite the fact that at the current moment, DMIND is able to generate only Christmas and COVID-19 short stories and poems imitating only Shakespeare’s writing style, in the future, these opportunities will be limitless. How cool is that, right? Imagine being able to transfer any famous writer’s style into any topic that you want! Well, that’s exactly what DMIND is all about!
