Stories from the Machine

Becca Rose, Cem Çakmak, Hugo Scurto. Hosted by Steffen Lepa

00:00

Watch on YouTube

Potato Computer Club

Becca Rose (Goldsmiths, University of London, UK)

Through my practice based PhD research I have developed the »Potato Computer Club« which is a space for affective relations with computers. The research explores the dialogue between deeply rooted feelings we have with ordinary objects (such as the potato) and new possibilities for technological futures. I design events and workshops as contexts for exploring the active relations between potatoes and computers. In these events participants have electrical encounters with potatoes. Through probing, sensing, and sonifying, they create new ways of listening, hearing, and seeing. As pedagogical apparatus, potatoes bring many modes of being and thinking into spaces of computing. They offer relational qualities that activate collective storytelling, and they become a vehicle for questioning computational mastery. Through imagining computation through the eye of the potato we enter a realm of different possibilities: slowness, storytelling, grounding, kinship, and collectiveness are some of the experiences participants say they have. The humble spud is a portal that opens up new meaning, memories, feelings, and humour with digital technology.

Becca Rose is an artist and PhD researcher at Goldsmiths, University of London. Potato Computer Club is part of their AHRC funded practice-based research exploring agency and material pedagogies in computing. Becca is co-curator of participatory computational arts festival Control Shift, and currently programming »Feeling Machines.« They are one third of electronic music making collective PRRRRRT! They currently teach physical computing on BA Computational Arts at Camberwell, University of Arts London, and previously on MA Design at Goldsmiths and BA Digital Media at UWE.

Implementing Ignorance and Unlearning Methods to AI Music Systems

Cem Çakmak (independent artist and researcher, TR/DE)

This research presentation will introduce concepts of Machine Unlearning and Artificial Ignorance, and discuss their musical potential musical applications. Many institutions around the world invest into musical creation through AI agents. While authenticity is not the focus of all research, oftentimes it is claimed that the AI is exhibiting some kind of creativity and originality. Unlearning and Ignorance are concepts in computer science that are concerned with privacy and data protection of individuals. This research aims to implement these concepts in open source AI music programs to intervene in converging processes. Thus, in terms of the CTM theme Portals, this presentation will be concerned with matters of data access and exclusion when training musical AI. Lastly, preliminary results from open-source musical AI programs that unlearn and ignore parts of their datasets will be presented for further discussion.

Cem Çakmak is a Sound Artist, Music Performer, Multimedia Designer and Researcher from Istanbul, based in Berlin since 2020. He often uses computer-assisted work that includes methods such as probability, improvisation, microtonality, errors, and binary logic in his artistic work. He holds a Masters degree in Sonic Arts and a PhD in Electronic Arts. As a researcher in spatial electronic music, he has given artist talks and paper presentations in conferences such as the Conference on Human Factors in Computing Systems (ACM-CHI), New Interfaces for Musical Expression (NIME), International Symposium on Electronic Arts (ISEA) and Web Audio Conference (WAC). His various academic research has been published in peer-reviewed journals such as the Journal of the Audio Engineering Society and Springer Series on Cultural Computing.

Soundwalking the Deepscape

Hugo Scurto (EUR ArTeC / Université Paris 8, FR)

Sounds produced by deep learning are gaining increased attention from musicians and artists, as well as from the general public. This global attention somewhat echoes the caring attention given by situated communities to their environments and the sounds those produce. Indeed, generative experiences of deep learning, including their streaming and spreading over social networks, are reminiscent of the multilayered soundscapes that continuously flow from the planet. Yet, despite progressive discourses on artificial intelligence (AI), deep learning perpetuates extractivism, as its industrial applications hegemonically exploits human and non-human sounds with high energy costs.

In this talk, I introduce the term »deepscape« to point at the global flows of sounds produced by deep learning, entangled with the human and material resources over which they are built. I describe ongoing collaborative radiophonic, virtual reality and installation works that attempt to reconnect our attention to the situated landscapes and soundscapes of the planet, by triggering soundwalk-inspired modes of listening to, and practicing with, deep learning. Rather than discussing AI opportunities and caveats for music or the creative industries, my talk seeks to open a portal to planetary concerns of AI, which I believe matter and deserve to be heard just as much.

Hugo Scurto is a researcher, musician and designer based in Marseille. Their research employs art, design and science to craft, prototype and diffract machine learning in an ecology of music. Their practice consists in creating, listening and performing with learning machines that reveal and reshape our musical entanglements with our environments. Hugo is currently postdoc at EUR ArTeC (Paris 8 / EnsadLab), and co-founding member of w.lfg.ng, a post-internet music collective. Before this, they completed a practice-based PhD in Machine Learning and Music Interaction at IRCAM, and were visiting student at the Department of Computing, Goldsmiths University of London.

Event Access

Please consult our Accessibility section for information on volume / lighting, mobility, age restrictions and other access information.