Počet záznamů: 1
From epistemology to ethics of deep networks
- 1.0577275 - FLÚ 2024 RIV GB eng C - Konferenční příspěvek (zahraniční konf.)
Hvorecký, Juraj
From epistemology to ethics of deep networks.
Proceedings of the AISB Convention 2023. Swansea: Swansea University, 2023 - (Müller, B.), s. 23-24. ISBN 978-1-908187-85-7.
[AISB Convention 2023. Swansea (GB), 13.04.2023-14.04.2023]
Grant CEP: GA ČR(CZ) GA20-14445S
Institucionální podpora: RVO:67985955
Klíčová slova: neural networks * adversarial attacks * epistemology * ethics * blackboxing
Obor OECD: Philosophy, History and Philosophy of science and technology
https://aisb.org.uk/wp-content/uploads/2023/05/aisb2023.pdf
Blame for various missteps of AI is often assigned to either programmers or users. The discovery adversarial attacks and the complex epistemology behind neural networks leads us to skepticism toward their capacity to work efficiently in domains where robust human categories are central. Ethics is one such a domain and AI systems might be particularly unsuitable for operations in this domain.
Trvalý link: https://hdl.handle.net/11104/0346996
Počet záznamů: 1