The connections between language, cognitive processes, community, expression and art have been pointed out always more often nowadays.
Though all the aforementioned domains seem to belong indisputably to the human nature, the breakthrough of the artificial intelligence, in particular the deep learning, has contributed to further belittle it, giving birth to algorithms enabling machines to discuss with people about any topic and produce remarkable art-works, including poetries and tales.
The communication pipeline which has been highlighted in the article “Plain terminology” flows in a different way when considering the human-machine interaction:
Human words-> Model-> Most probable words
Its words-> Its concepts
So the natural tendency to believe that concepts are lurking behind words persists even in case we are speaking with a machine, regardless of our general awareness of the very different kind of processes which underlie the automatic text generation. This is called the Eliza effect.
Eliza is the first chatbot ever created.
Despite the computational and technological limitation of the 70´s, the program´s quality could fool many people, including people working at the project.
If more than the half of tested people believe to have talked with humans (or, in general, with entities provided with consciousness) than we can say that the machine (or the program) has successfully passed the “Turing test”.
In the last decade the amount of softwares able to pass this test has increased dramatically. For this reason and for its importance in our era, as well as in our daily life, in this and in the next episodes of this mini-series dedicated to the computational creativity we will explore some of the methods and results of bold computational linguists attempting to make machines more similar to us.
In the following video Pablo Gervas explains what inspired him to dedicate his efforts to this cause, and what has been already achieved until now.
You can access the complete version of this video at the following link:
If you wish to deepen your understanding of tale-generation algorithms click on the following links:
Selected and described by Cosimo Palma, Communication Trainee at the Terminology Coordination Unit of the European Parliament (Luxembourg).