Scholarlink

<p>Believing that something has not happened because we have not heard about it is a common form of reasoning in everyday human life. Drawing on the work of Goldberg (2010) and Pedersen and Kallestrup (2013), who provide a systematic epistemological account and formalization of inference from epistemic absence, I examine whether the notion of epistemic coverage (EC), which underpins the reliability of absence-based inference (ABI), can be directly applied to algorithmic decision-making (ADM). I argue that although there are cases of reliable algorithmic ABI in a purely epistemic sense, ABI remains fundamentally a human way of reasoning, as it involves relying on one’s community for epistemic coverage. This paper problematizes the application of ABI to ADM and offers an initial approach to how ABI might be implemented algorithmically.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

On Reliable Algorithmic Absence-Based Inference

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年7月10日

查看全文
  • Ekaterina Pashevich

摘要

Believing that something has not happened because we have not heard about it is a common form of reasoning in everyday human life. Drawing on the work of Goldberg (2010) and Pedersen and Kallestrup (2013), who provide a systematic epistemological account and formalization of inference from epistemic absence, I examine whether the notion of epistemic coverage (EC), which underpins the reliability of absence-based inference (ABI), can be directly applied to algorithmic decision-making (ADM). I argue that although there are cases of reliable algorithmic ABI in a purely epistemic sense, ABI remains fundamentally a human way of reasoning, as it involves relying on one’s community for epistemic coverage. This paper problematizes the application of ABI to ADM and offers an initial approach to how ABI might be implemented algorithmically.