Scholarlink

<p>Since at least the 1990s, social and media studies contributed to presenting algorithms as human, cultural constructs. More recently though, fine-grained ethnographic inquiries went further in the analysis and showed that most algorithms derive from benchmark referential datasets – often called “ground truth” – that gather input data and output targets, thereby establishing what can be approximated computationally and evaluated statistically. In this commentary, I briefly review the most recent results on this line of research one may call ground truth studies. I then consider some objections to this ground truth-centered conception of algorithms and point out avenues for future thinking.</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

Examining Algorithms in the Light of their Ground Truth Datasets: Results, Objections, and Avenues of Reflection

  • 期刊论文

  • 发表日期：

    2025年7月11日

查看全文
  • Florian Jaton

摘要

Since at least the 1990s, social and media studies contributed to presenting algorithms as human, cultural constructs. More recently though, fine-grained ethnographic inquiries went further in the analysis and showed that most algorithms derive from benchmark referential datasets – often called “ground truth” – that gather input data and output targets, thereby establishing what can be approximated computationally and evaluated statistically. In this commentary, I briefly review the most recent results on this line of research one may call ground truth studies. I then consider some objections to this ground truth-centered conception of algorithms and point out avenues for future thinking.