Entangling Umwelten: Towards a New Ecology of Machine Learning, Moving Images and Nonhuman Worlds

这个以实践为基础的博士学位将是第一个扩展的学术研究，以探索生物学家雅各布·冯·于克斯库尔的Umwelt概念在当代电影和移动图像的理论和实践中的应用。在他1934年的作品《动物和人类世界的探索》中，Uexkull将Umwelt定义为每个动物和人类的感知和有意义的环境的“周围世界”，强调每个生物都决定了自己周围环境的意义。我将遵循Inga Pollmann先前在电影活力论（2018）中的理论化论点，即Umwelt理论源于Uexkull对早期电影摄影的实验，并且它可以作为一种使传统电影话语复杂化的方式，作为一种作用于而不是与存在的环境。我将探讨Umwelt理论如何在政治上和智力上激进，如果我们考虑Uexkull之后的思想家和艺术家如何使用它来创造新的方式来改变非人类中心的凝视和与非人类在他们的环境中。这将通过机器学习装置和电影的制作以及扩展thems.The论文将探讨当代电影和电影作品的选择，通过透镜的关键领域，建立在Umwelt理论在整个世纪后期到21世纪。对于博士的实践层面，我将使用尖端的机器学习模型“训练”动物运动摄影档案，并创建非人类运动图像“解释”这些档案，我称之为“后人类运动”。这一实用的系列将通过一系列关于人类与动物相遇的短片来加强，这些短片采用了各种机器学习方法，例如“升级”算法，作为电影语法的新的、实验性的和激进的形式。

This practice-based PhD will be the first extended academic study to explore applications of the biologist Jakob von Uexkull's concept of Umwelt to the theory and practice of contemporary film and moving images. In his work A Foray into the Worlds of Animals and Humans (1934), Uexkull defines Umwelt as the 'surround-world' of each animal's and human's perceptual and meaningful environment, emphasising that each living being determines the meanings in their own surroundings. I will follow Inga Pollmann's previously under-theorised argument in Cinematic Vitalism (2018) that Umwelt theory grew out of Uexkull's experiments with early cinematography, and that it can stand as a way of complicating conventional discourses of cinema as a milieu which acts upon rather than with being. I will explore how Umwelt theory can be politically and intellectually radical if we consider how thinkers and artists after Uexkull have used it to create openings for new ways of altering the anthropocentric gaze and being with nonhumans in their environments. This will be done through the production of machine learning installations and films alongside an extended thesis.The thesis will explore a selection of contemporary film and moving image works through the lens of critical fields which build upon Umwelt theory across the later 20th century and into the 21st. For the PhD's practical dimension, I will produce a collection of videos under the title Metabolisms using cutting-edge machine learning models 'trained' on archives of animal locomotion photography, and which create nonhuman moving image 'interpretations' of those archives which I call 'posthuman locomotion'. This practical strand will be enhanced by a series of short essay films about human-animal encounters which employ various machine learning methodologies, such as 'upscaling' algorithms, as new, experimental and radical forms of cinematic grammar.