Savic, Selena

Profilbild
E-Mail-Adresse
Geburtsdatum
Projekt
Organisationseinheiten
Berufsbeschreibung
Nachname
Savic
Vorname
Selena
Name
Savic, Selena

Filter

2020 - 20233
Filter zurücksetzen

Einstellungen

Institut: Institut Experimentelles Design und Medienkulturen × Thema: machine learning ×

Suchergebnisse

Gerade angezeigt 1 - 4 von 4
  • Vorschaubild
    Publikation
    Making Arguments with Data: Resisting Appropriation and Assumption of Access / Reason in Machine Learning Training Processes
    (Weizenbaum Institute for the Networked Society, 30.10.2023) Savic, Selena; Martins, Yann Patrick [in: Weizenbaum Journal of the Digital Society]
    This article presents an approach to practicing ethics when working with large datasets and designing data representations. Inspired by feminist critique of technoscience and recent problematizations of digital literacy, we argue that machine learning models can be navigated in a multi-narrative manner when access to training data is well articulated and understood. We programmed and used web-based interfaces to sort, organize, and explore a community-run digital archive of radio signals. An additional perspective on the question of working with datasets is offered from the experience of teaching image synthesis with freely accessible online tools. We hold that the main challenge to social transformations related to digital technologies comes from lingering forms of colonialism and extractive relationships that easily move in and out of the digital domain. To counter both the unfounded narratives of techno-optimismand the universalizing critique of technology, we discuss an approachto data and networks that enables a situated critique of datafication and correlationism from within.
    01A - Beitrag in wissenschaftlicher Zeitschrift
  • Vorschaubild nicht verfügbar
    Publikation
    Making Arguments with Data
    (09.06.2022) Savic, Selena; Martins, Yann Patrick
    Whether we are discussing measures in order to ‘flatten the curve’ in the ongoing pandemic, or what to wear in face of the most recent weather forecast, we make arguments based on patterns and trends observed in data. What makes these patterns observable? Making arguments with data requires critical engagement with datasets, as well as computational processes to gather data, to organize and model their relationships. This article presents an approach to practicing ethics when working with large datasets and designing data representations. The arguments we make are based on the development and use of a computational instrument, and working with digital archives. We programmed and used web-based interfaces to sort, organize and explore a community-ran archive of radio signals. Inspired by feminist critique of technoscience and recent problematizations of digital literacy, we argue that one can navigate machine learning models in a multi-narrative manner, and that knowledge of radio signals or any other technical artefact transgresses domains. We propose visual explorations of complex data structures that enable storytelling and an understanding of datasets that resists extraction of discrete identities from the data. We hold that the main challenge to sovereignty comes from lingering forms of colonialism and extractive relationships that easily move in and out of the digital domain. Countering both the unbased narratives of techno-optimism, and the universalizing critique of technology, we discuss an approach to data and networks that enables a situated critique of datafication and correlationism from within.
    06 - Präsentation
  • Vorschaubild
    Publikation
    Architectonic Studies of Radio Signals: Reorganizing Archives of Data/Natures In Their Own Terms
    (18.08.2020) Savic, Selena
    As we slowly accustom to thinking about planetary issues through the notion of ‘assemblage’ rather than that of the ‘system’, we get better at acknowledging complex entanglements between living and inert, between social and technical. This paper presents a critical reflection on the use of machine learning techniques to support reasoning about natural phenomena. It engages data/natures by focusing on data radio signals: a phenomenon that pertains to both culture (telecommunications) and nature (atmospheric lightning discharges). Signal Identification Guide Wiki, a rich archive of signals observed and documented by a community of radio enthusiasts is the starting point of this study. In order to articulate alternative ways to study and engage with radio signals, I develop 'digital observatories': new methods for organizing and navigating abundant digital information based on critical use of self-organising map algorithm. I present a study of distribution patterns and clustering of signal qualities, when signals are reduced to spectrograms (visual representation of signal frequency composition). This 'digital observatory' aims to facilitate speculation on the connection between signal representation and technical communication protocols, by enabling the observer to identify criteria of similarity, and intervene in this organised space by adding new (real or imaginary) data. The project contributes to the fields of STS and experimental design research with an interest in the digital, unsettling the dichotomies previously described and providing avenues for recognition of the entangled nature of matter and information, of human and other-than-human, beyond simple ontological distinctions.
    06 - Präsentation
  • Logo des Projekt