Computational Ecologies

Andrew Dwyer, University of Durham
:

Malicious software and other computational materialities have rarely been understood as more-than-human, but rather as technologies through post-humanism and beyond (Hui, 2019). In this paper, I wish to extend and complicate this through autoethnographic fieldwork to query computation as more-than-human. Computational ecologies and terrains are places brimming with agency, choices, and decisions that require further attention in geography and beyond. Biological and computational viruses have exemplified a liminality between (non-)life, inspiring cybernetics to flatten distinctions through abstraction to information exchange. Using N. Katherine Hayles’ various works (2019, 2017), I critique this reductive comparison to argue that computation, across various terrains and ecologies, enable ‘autonomous’ choices to be made that open up the political not only to computation, but also to animals and plants through the concept of ‘choice-making’. That is, a wider political ecology whereby materials exchange, interpret, and act upon signs. I then use this to complicate more-than-human ‘digital’ ecologies to incorporate various computational material-virtual terrains incorporating hardware, the software patch, as well as the light across silica of fibre optic cables. This then permits me to initiate thinking on the complex interactions and implications of and between (non-)computational ecologies.

Andrew Dwyer is an Addison Wheeler Research Fellow at Durham University, UK. His research spans cybersecurity, malware, computational agency, as well as the role of machine learning algorithms in the production and negotiation of security. He completed his PhD – Malware Ecologies – at the University of Oxford in 2019.

Katherine Dow, University of Cambridge
Sophia Doyle, Goldsmiths University of London
:

In 2020, London Freedom Seed Bank (LFSB), a group of seed-savers and seed activists based in London, decided to use Airtable to create a collaborative seed database, the LFSBase. The initial purpose was to establish an open-access database of the Bank’s seed collection, to facilitate saving and sharing of seeds amongst their growers’ network; it has since expanded to capture data about seed-saving in London more generally. Not only has the database proven a very effective tool for organising, issue-tracking, recording and archiving the material seed stock, it has also become a virtual space for recording and engaging with growers’ encounters with particular seeds and varieties, growing processes and localised conditions. LFSBase is envisioned as a digitally mediated space in which the largely invisible network of human growers, plants, pollinators, soil and climate spanning the city of London is made tangible. As a co-director and a member of LFSB’s steering group, as well as researchers, we have observed and participated in this project since its inception. In this paper, we will present the LFSBase, to illustrate some of the broader themes of knowledge, mobility, sharing, generation and time that this digital project, and its articulation by its creators, reveals. In particular, we will discuss how concepts of variety, genetic purity and local adaptability are flexibly put to work in the database. What model(s) of seed, people and place does LFSBase produce and reproduce when low-tech seed stories come into contact with a state-of-the-art digital platform?

Katharine Dow is a senior research associate and deputy director of the Reproductive Sociology Research Group (ReproSoc) at the University of Cambridge. She specialises in connections between reproductive and environmental concerns and activism, from a multispecies perspective. She is the author of Making a Good Life (Princeton University Press, 2016). She is also on the steering group of London Freedom Seed Bank.

Sophia Doyle is a postgraduate student at Goldsmiths, studying the MA Postcolonial Culture and Global Policy. Her research focuses on critical ecologies, seed sovereignty, food and land justice and folk knowledges. She is co-director of the London Freedom Seed Bank and is writing her dissertation on the Seed Bank’s data project and how it relates to issues of knowledge commoning, open access and grower-seed relationships. 

Myung-Ae Choi, Center for Anthropocene Studies, KAIST
Seunghwa Yoo, National Institute of Ecology
Junyoung Byun, School of Electrical Engineering, KAIST
Hyojun Koh, School of Electrical Engineering, KAIST
Byungjun Park, School of Electrical Engineering, KAIST
Jeongsoo Kim, School of Electrical Engineering, KAIST
Changick Kim, School of Electrical Engineering, KAIST:

Recent innovations in digital technologies have dramatically shifted the ways in which nature conservation is thought about and practised. This study focuses on the use of artificial intelligence (AI) for environmental research by developing deep-learning algorithms for wildlife monitoring; as well as developing a relational analysis of it. This interdisciplinary research project has two aims. First, we develop computer vision systems for analysing wildlife images from ecologists and trap cameras in and around the Korean Demilitarized Zone (DMZ) – which we call the “AI ecologist”. The algorithms can perform automated species identification of a dozen of mammal species and several crane species, while making population-count estimations of selected crane species. Second, by drawing on our own experience of developing deep-learning based vision systems, we explore the ways in which the machine and the human intersect in the formation and operation of AI. We specifically focus on the epistemologies of AI ecologists and human ecologists by examining their overlapping and diverging ways of making sense of particular animal species. This involves discussions of translation, agency, and friction, which opens up the space for surprises and creativity. In doing so, instead of making recourse to the prevalent ideas of AI as an autonomous system or an inert device under human control, this study develops an understanding of AI as relational achievement that emerges from the performance multispecies entanglement.

Myung-Ae Choi is an environmental geographer exploring the politics, cultures, geographies, and technologies of conservation in East Asian developmental state, especially South Korea. She is Research Assistant Professor at the Centre of Anthropocene Studies in Korea Advanced Institute of Science and Technology (KAIST).

Helge Peters, University of Oxford
Nathanael Sheehan, University College London:

In recent years inland surface water quality in England has become a matter of public controversy. Drawing on materialist and relational theories of the public in more-than-human geography and STS, we present a computational approach to the critical and reflexive study of the way in which environmental publics digitally mediate their ecological matters of concern. Using machine learning, natural language processing, and interactive web mapping, we explore the metric and participatory affordances of social media platform Twitter in order to interrogate, and invite contribution to, the performance of water as a controversial environmental issue. We present three main findings: a) Topic modelling of large data sets both aids mapping the way in which the multiple ontologies of water are digitally mediated and critically distinguishing between place-based and professional performances of water as a matter of concern. b) Complementing an interpretive approach to textual analysis, sentiment analysis provides a quantitative lens on the performance of digital publics critical of environmental management when social media users attempt to enact water as a matter of ecological quality rather than flood risk. c) Interactive web mapping of sentiment about London water invites reflection on more reflexive and participatory relations between environmental social science and its publics when studying public controversy from a relational and performative perspective. We discuss the potentials and limitations of the computational methods generating these three findings for more-than-human geographies of environmental publics.

Helge Peters is a postdoctoral researcher at the School of Geography and the Environment, University of Oxford. He currently conducts transdisciplinary research on urban water management in London using ethnographic and computational methods. Helge holds a doctorate in geography from Oxford and was a fellow at the Harvard Kennedy School. Prior to that he obtained degrees in media and communications from Goldsmiths College and the Berlin University of the Arts. He lives in Berlin.

Nathanael Sheehan’s research focuses on geo-computation methods for sustainable urban planning. Since graduating from his MSc at CASA UCL, he has worked on a UKRI funded project and contributed to several open-source R packages.