The reflection made by intelligent agents, which means the entities trained by machine learning algorithms, was defined as unpredictable one. According to the human-computer interaction before the emergent of AI, people also input commands and let machines react people’s actions. However, the difference is that people can predict what kind of reaction the machine can make. I was thinking if the randomness data behind intelligent agents can be used to analyze what the AI bring to us.
“The fact of being done, chosen, etc. without somebody deciding in advance what is going to happen, or without any regular pattern.”
-----Oxford Learner’s Dictionaries
Randomness is a kind of tricky stuff. The definition of randomness in the dictionary describe it as a fact, which is a status before people proving it. However, In scientific research field, it seems like people have not found a real mechanism which can prove that if randomness exists in this world, but we do have a vocabulary to describe this kind of phenomena. As so far, randomness still can not to be real randomness, the only thing we can do is to modify the level of its complexity, which Verbeeck called pseudo-random (Verbeeck). For example, technically, randomness used in both creative coding, such as noise and random functions, and AI algorithms, such as generative adversarial network (GAN), aims to get unpredictable values, but the difference is in the algorithms, to what extent human would like to control the so called random outcomes directly. To prove the existence of randomness is contradiction. In common cases, we usually put it in a context with some limitations or rules, when we talk about randomness(Landsman). For example, we see the randomness only in one thousand numbers, which is a finite condition. Figure 1 shows the visual effects the reflect the brightness value of an images.
(figure 1)
Ref:
Landsman, Klaas, and Ellen van Wolde, editors. The Challenge of Chance: A Multidisciplinary Approach from Science and the Humanities. Springer International Publishing, 2016. DOI.org (Crossref), doi:10.1007/978-3-319-26300-7.
Verbeeck, Kenny. Randomness as a Generative Principle in Art and Architecture. p. 97.
Comments