"DeepAesthetics is a study of the aesthetics of machine learning and an account of the operations of AI and data science. Drawing on Deleuzian theory, Anna Munster thinks about how machine learning itself is a complex sociotechnical system-for example, she links anti-Black racism in AI to complex relations of statistical techniques derived from eugenics. Munster provides concepts and proposes an approach for probing experience reconfigured by machine learning. 'Deep aesthetics' characterizes AI as it is figured by statistical, platform, and quantifying tendencies, and draws on deep learning-the subfield of machine learning that uses neural network architectures. Situated across critical AI studies, software studies, science and technology studies, and art theory, the project of DeepAesthetics is to shift focus from quantifying computational 'experience' to how computation qualifies. DeepAesthetics offers a way to make sense of the insensible and frequently imperceptible forms of nonlinear and continuously modulating statistical function and calculation. Alighting on the domains and operations of image production; statistical racialization; AI conversational agents; and critical AI art, Munster analyzes how machine learning is operationally entangled withracialized, neurotypical, and cognitivist modes of producing knowledge and experience"--
Anna Munster explores aesthetics, artificial intelligence, and machine learning to understand the contours of computational experience and the possibility to artfully use AI to create new futures.
Computation has now been reconfigured by machine learning: those technical processes and operations that yoke together statistics and computer science to create artificial intelligence (AI) by furnishing vast datasets to learn tasks and predict outcomes. In DeepAesthetics, Anna Munster examines the range of more-than-human experiences this transformation has engendered and considers how those experiences can be qualitative as well as quantitative. Drawing on process philosophy, Munster approaches computational experience through its relations and operations. She combines deep learning—the subfield of machine learning that uses neural network architectures—and aesthetics to offer a way to understand the insensible and frequently imperceptible forms of nonlinear and continuously modulating statistical function. Attending to the domains and operations of image production, statistical racialization, AI conversational agents, and critical AI art, Munster analyzes how machine learning is operationally entangled with racialized, neurotypical, and cognitivist modes of producing knowledge and experience. She approaches machine learning as events through which a different sensibility registers, one in which AI is populated by oddness, disjunctions, and surprises, and where artful engagement with machine learning fosters indeterminate futures.
