The Lion and the Bull: a methodology of embracing failure (i, ii, iii)

A lecture-performance

Debashis Sinha | Toronto Metropolitan University

Debashis Sinha: bio

Debashis Sinha’s research-creation projects entangle sonic arts, ethnographic practice, music composition and urban studies, with numerous exhibitions and releases to his credit. His research in crafting speculative mythologies  using machine learning and AI is targeted at uncovering new strategies of story creation, and has been presented in a number of artistic venues and contexts as well as international conferences held by NeurIPS, the Royal Anthropological Institute, Carleton University, NoiseFloor UK, and others.

He is currently an assistant professor at Toronto Metropolitan University.

Previous iterations of the research:

This will be a lecture performance version of an audio paper that was part of the (de)Stabilizing Diffusions public exhibition @ SAT Montreal, May 15-25th, 2023, presented by Milieux Institute, SAT Montreal, and MUTEK. This artistic research exhibition was attached to the unStable Diffusions conference on AI, presented by Milieux Institute. Each part of this audio paper was presented on a separate iPad with headphones. From the exhibition notes: “The exhibition (de)Stabilizing Diffusions aims to defy the hype around generative AI tools by reshifting the focus on investigating these systems in and through artistic practice.“

Thanks to Maurice Jones of the the Milieux Institute for the invitation.

Additional information on the presentation format:

The lecture performance of this material is not a linear rebroadcast of the audio paper. The text, spoken aloud, will be accompanied by a playback/remix of the other audio material in the original paper, remixed and performed/improvised to varying degrees. Drawing on performance, improvisation, lecture, DJ and electronic music tactics, the presentation will be a unique iteration of the research belonging to the specific performance at the conference. This performance is part of a larger imperative in my research to highlight and uplift modes of knowledge exchange connected to relationality, inspiration and improvisation, and to connect research-creation methods with processes of decolonization.

Presentation Image:

Abstract:

The process of equating algorithmic tools as “complete” repositories of knowledge continues apace, with predictable results. The rejection of slower, more traditional modes of engagement in favour of quantifiable data easily parsed by mathematical algorithms has resulted in prioritizing archives that prioritize a narrow range of human data optimized for use by algorithmic tools, data that encapsulates many of the failures of human society. The inevitable result of the use and priority of this data alongside the unquestioning acceptance of the imperatives of machine learning - speed, accuracy, scalability - arguably replicates many of the foibles of our history as a species. Cultural practice in general is a slippery but deeply felt affair, often communicating the “underside of meaning” (from Barthes, 1976). In algorithmic times this communication focus can participate in the many cautions of AI and machine learning already extant in activist circles.

In this lecture-performance, Sinha illustrates the process of refusing to rely on the traditional imperatives of classification and reproduction in machine learning by presenting a condensed composition process for the work “The Lion and The Bull” from his recent release “Adeva_v000_04”. The ideas considered are offered both as a method of sound generation and as a possible strategy of refusing the status quo as regards algorithmic tools, instead leveraging their inevitable failures as a deliberate strategy of synthesis of slow human modes of storytelling with fast algortihmic outputs, embracing the inaccuracies and fuzziness of both processes to uplift models of cultural possibilities and community building.

contact: sinhad@torontomu.ca

Other information (please scroll):

Presentation format: 

Short presentation (20m/20m); Lecture performance, with sound and slides. 

 

Spatial and/or technical requirements:

Proper sound support essential - stereo sound and projection for a Google Slide presentation required. 

120v AC power if possible

Enough table space for laptop, MIDI controller and sound card (medium sized desk will do, please confirm with presenter)

 

Potential risks or ethical concerns:

None noted.

 

 

 

part i (audio paper version)

stage plot / technical info