Raffaele Andrea BUONO

I have recently completed a PhD at University College London (UCL), under the supervision of Prof. Ludovic Coupaye and Dr. Tone Walford. I have conducted research in a Japanese robotics and AI laboratory, where I have attempted to delineate a theory of technicity for the machine learning models being tested in my fieldsite. In analysing the inferential, meaning-making and interactional profiles of unsupervised learning techniques, variational autoencoding architectures, and Bayesian sampling methods, my research pinpointed to the difficulties algorithmic models face in operating through the generative transductions and abductions at the basis of human modes of individuation, and which these models seek to replicate. While conducting my doctoral research, I have also been exposed to many of the neuroscientific theories, models and frameworks underpinning both theories of mind and cognition, and their engineering transpositions. It is this latter line that my current research will be focusing on, as a Postdoctoral Fellow within Prof. Matteo Pasquinelli's ERC AIMODELS project. I am currently investigating, historically and ethnographically, the emergence of unsupervised learning models as efficient algorithmic approximations of cognitive processes, and how their efficiency in turn delineates specific ontological and epistemic contours to human models of cognition, ultimately coalescing into paradigms such as predictive processing and the Free Energy Principle. I am particularly interested in (1) how algorithmic understandings of modelling travel across and between machine learning research and computational neuroscience; (2) the kinds of visions of life, sociality and futurity such visions concretise; (3) what gets to be left aside. My research is particularly indebted by the works of Gilbert Simondon and Charles Sanders Peirce, as well as the 'enactive turn' I consider they anticipated in many ways; as part of my current work, I am particularly interested in further developing a thoroughline between these two thinkers, particularly by triangulating their ideas around transduction, abduction, semiosis and individuation. Currently, I am working on two monographs (one on transduction, and one on the emergence of predictive processing paradigms), and a bunch of articles related to questions of prediction, generativity, modelling and unsupervised learning. I believe in the value of discussion, and I am always open to writing collaboratively, particularly with colleagues coming from different disciplines. In this sense, I welcome collaborations with engineers and HCI specialists, as I have in the past few years through my work with Prof. Nadia Bianchi-Berthouze on developing different models for understanding (and engineering) human-robot relations via touch.