为什么：关于因果关系的新科学
The Book of Why
Judea Pearl、Dana Mackenzie · 初版 2018 年
59人推荐
103位读者提及
144次提及
为什么值得读
- 系统提炼了统计与AI巨匠在因果推断上的成果
- 点醒读者大多数机器学习本质只是曲线拟合
- 以吸烟危害研究为例说明如何厘清因果关系
一位荣获 Turing Award 的计算机科学家兼统计学家阐述了理解因果关系如何彻底变革了科学，并将颠覆人工智能。“相关性不等于因果关系。”一个多世纪以来，科学家们一直将这句格言奉为圭臬，甚至导致因果讨论在事实上被全面封禁。如今，这一禁忌已被打破。由 Judea Pearl 及其同事发起的因果革命打破了一个世纪的迷思，并将因果关系——对原因与结果的研究——建立在了坚实的科学基础之上。他的工作解释了我们如何能够探知那些看似简单的事情，比如是否下过雨
Hacker News 上怎么说
正在加载完整评论……
…《The Soul of a New Machine》，或者如果你喜欢看《Halt and Catch Fire》（奔腾年代），你很可能也会喜欢《Dreaming in Code》。受 Marty Cagan 的启发。对产品管理的核心要领做了非常扎实的综述。Judea Pearl 的《The Book of Why: The New Science of Cause and Effect》。Judea Pearl 当然是统计学和 AI 领域的泰斗，这本书提炼了他在“因果推断”（causal inference）方面的研究成果，并将其条理清晰地展现在…
…The Soul of a New Machine, or if you like watching Halt and Catch Fire, you may well like Dreaming in Code. Inspired by Marty Cagan. Really solid overview of the essentials of product management. The Book of Why: The New Science of Cause and Effect by Judea Pearl. Judea Pearl is, of course, a giant in the worlds of statistics and AI, and this book distills his work on "causal inference" and lays it all out in…
我猜这篇被疯狂点踩是因为提到了兰德（Rand）。虽然我也不怎么感冒她，但这也太有失公允了：1. 《梦断代码》（Dreaming in Code）和《为什么》（The Book of Why）都很出色（另外两本我没读过）；2. 哪怕我不认同兰德的整套世界观，她提出的一些观点也是某些人有必要了解的。
I presume this was heavily downvoted because of the Rand. While I'm not a fan of her either, that's unfair: 1. Dreaming in Code and The Book of Why are excellent (I haven't read the other two), and 2. Rand makes some points that some people need to learn, even if I don't buy her whole worldview.
可悲的是，社交媒体成了新时代的吸烟。我鼓励每个人都去读读 Judea Pearl 的《The Book of Why》。它揭露了当年人们是如何利用所谓的“研究”和“调查”来对吸烟的危害散布怀疑的。我们现在对待社交媒体也是这种处境，但是，我们现在有了像因果推断（causal inference）这样的新工具…
Sadly, social media is the new smoking. I encourage everyone to read "the book of why" by Judea Pearl. It illustrates how "research" and "studies" were used to sew doubt about the dangers of smoking. That's where we're at with social media, BUT, we have new tools like causal inference…
没错，Judea Pearl 的《The Book of Why》让我大开眼界，意识到机器学习里的绝大部分操作说到底其实就是曲线拟合（curve fitting）。这让我想起了 Chomsky 说过的一句话，大意是说通过拍摄录像来试图总结物理定律…
exactly, Judea Pearl's The Book of Why opened my eyes to the fact that most of what happens in machine learning is really just curve fitting It connected with what i've heard Chomsky say about trying to develop laws of physics by filming…
…一条立足于你现状的清晰路径。对我来说，Gelman 的《Data Analysis Using Regression and Multilevel/Hierarchical Models》[0] 把许多许多观点讲得非常透彻。最近，我隐隐觉得/期望 Pearl 的《The Book of Why》[1] 能把这提升到一个全新的高度。[0] http://www.stat.columbia.edu/~gelman/arm/ [1] https://www.amazon.com/Book-Why-Science-Cause-Effect/dp/0465...
…a clear path from where you are. For me Gelmans Data Analysis Using Regression and Multilevel/Hierarchical Models [0] drove home many, many points. More recently, I have a sense/hope that Pearl's The Book of Why [1] might take this to yet another level. [0] http://www.stat.columbia.edu/~gelman/arm/ [1] https://www.amazon.com/Book-Why-Science-Cause-Effect/dp/0465...
…没有心智模型去思考“如果……会怎样”（what if），也无法利用反事实（counter-factuals）去推敲如果没做某事会发生什么。目前它说白了就是被美化了的模式匹配（pattern matching）。读了 Judea Pearl 的《The Book of Why》之后，我对 AI 的看法确实清醒理智了许多。
…have a mental model to ponder 'what if' and can't use counter-factuals to ponder what would happen if it didn't do something. It's just glorified pattern matching currently. Reading Judea Pearl's The Book Of Why certainly sobered my outlook on AI.