The last decade has seen the triumphant march of machine learning techniques, and the knowledge of their practical utility is now widespread even among laypeople. However, in the case of neuroscience, these methods are more than just another convenient tool. Their ability to solve a wide range of complex problems at a level comparable to humans and their low operating costs have made them the new preferred model for testing hypotheses about the functioning of biological brains. However, such an approach carries special risks – seemingly insignificant engineering decisions can, as it turns out, radically alter the results obtained in experiments.
During the lecture, we will learn why artificial neurons sometimes behave confusingly similar to real ones (and what crabs have to do with it) and pay attention to the bitter lesson related to the search for digital equivalents of living reticular cells responsible for orientation in space.
Radosław Łazarz, MSc Eng.
Employee of the Institute of Computer Science at the Faculty of Computer Science, Electronics, and Communications at the AGH University. In his works, he has been focused on automatic structural pattern processing and multi-criteria optimisation. He believes that (paraphrasing Clarke's third law, which is well-known in pop culture) if we allow almost no one to know how a given technology works, we can be sure that soon everyone else will start thinking of it in magical terms. To at least dismiss this unpleasant prospect to some extent, he has been trying to tell stories from his discipline in a way that makes them (relatively) painless to listen to for almost a decade now.
Reports of advances in artificial intelligence are constructing new mythologies of machines. In the conversation about the limits of technology progression, there are accelerationist positions – fantasies of mind transfer to the virtual world and stories about human beings improved by fusing with AI. On the other hand, the mythologies of the technical sphere are constantly being tested and hacked by those who skilfully navigate the system and find its cracks, revealing the underlying principles of its operation and optimisation.
The lecture will focus on tactics for reconfiguring the mythology of machines, projects in which artificial intelligence can appear in many, often unexpected, variations.
Dr Anna Olszewska
Assistant professor at the Faculty of Humanities at the AGH University of Krakow. Manages the Re:Senster programme which aims to reactivate cybernetic sculpture by Edward Ihnatowicz and co-conducts research within EduVRLab and Department of Studies on Culture and Research on Digital Era. Her research interests concern mostly science and technologies as well as the history of art, with particular emphasis on the ontology of artificial vision and post-growth technologies.
Besides her work at the AGH University, Dr Olszewska is also employed in the Graphic Collections Department in the Scientific Library of the PAAS and the PAS in Cracow. As a curator and researcher, she collaborated with numerous cultural institutions, including: National Museum in Krakow, Foundation for Visual Arts, WRO Art Center, Tranzit.ro, Olomouc Museum of Art, Center for Urban History of East Central Europe Building in Lviv, and Sapporo International Art Festival.