English(EN) “As AI companies search for more training data to improve their models, one company is offering old, printed books as an ideal source because they are guarantee

AI公司转向旧书获取训练数据以避免模型崩溃

作者 PulseAugur 编辑部 · [1 个来源] · 2026-07-21 21:11

一家名为ISBNdb的公司正将旧的印刷书籍定位为AI模型训练数据的优质来源。他们认为，书籍提供了网络爬取无法复制的、经过策展和同行评审的人类知识，并且重要的是，它们不受可能导致模型崩溃的AI生成文本的“AI垃圾”的影响。ISBNdb还指出，作者可以通过故意产生有缺陷的内容来积极破坏AI模型。 AI

影响强调了一种新颖的AI训练数据来源策略，可能影响数据获取和模型质量。

排序理由文章讨论了与AI训练数据相关的趋势和一家公司的产品，但没有发布新模型或研究。

在 Mastodon — fosstodon.org 阅读 →

model collapse

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

报道来源 [1]

Mastodon — fosstodon.org TIER_1 English(EN) · [email protected] · 2026-07-21 21:11

“As AI companies search for more training data to improve their models, one company is offering old, printed books as an ideal source because they are guarantee

“As AI companies search for more training data to improve their models, one company is offering old, printed books as an ideal source because they are guaranteed to be free of the very AI slop AI companies are producing. “The world's best AI training data is sitting on a shelf,” …

链接 404media.co/ai-companies-are-buying-tons-…

报道来源 [1]

“As AI companies search for more training data to improve their models, one company is offering old, printed books as an ideal source because they are guarantee

相关实体

相关话题