The Book of Why

来自集智百科 - 复杂系统|人工智能|复杂科学|复杂网络|自组织
Wits讨论 | 贡献2022年5月10日 (二) 11:42的版本
跳到导航 跳到搜索
  • 此词条由因果科学读书会词条梳理志愿者雨轩翻译审校,未经专家审核,带来阅读不便,请见谅。
  • 第一段英文为原维基百科的内容,第二段英文为摘除多余格式的英文原文,方便大家审校,以及第三段为机器翻译的中文。
Authors Judea Pearl and Dana Mackenzie
Language English
Subjects Causality, Causal Inference, Statistics
Publisher Basic Books (US)Penguin (UK)
Publication date 2018
ISBN 9780141982410
Preceded by Causal Inference in Statistics: A Primer

《为什么:因果新科学》是计算机科学家朱迪亚·珀尔和作家达娜·麦肯齐在2018年合著的非小说类书籍。这本书为普通读者从统计和哲学的观点探讨了因果关系和因果推理的问题。

目录

  • 引言:思维胜于数据
  • 第一章:因果关系之梯
  • 第二章:从海盗到豚鼠:因果推断的起源
  • 第三章:从证据到原因:当贝叶斯牧师遇见福尔摩斯先生
  • 第四章:混杂和去混杂:或者,消灭潜伏变量
  • 第五章:烟雾缭绕的争论:消除迷雾,澄清事实
  • 第六章:大量的悖论!
  • 第七章:超越统计调整:征服干预之峰
  • 第八章:反事实:探索关于“假如”的世界
  • 第九章:中介:寻找隐藏的作用机制
  • 第十章:大数据,人工智能和重要问题

致谢

参考文献

本书包含了10个章节和引言部分。

引言:思维胜于数据

导言描述了20世纪早期统计方法在陈述变量之间的因果关系方面的不足。在此之后,作者描述了他们称之为“因果革命”的事件,它始于20世纪中期,为描述因果关系提供了新的概念和数学工具。

第一章:因果关系之梯

第一章介绍了“因果阶梯”——一个用来说明因果推理三个层次的图表。第一级是“关联”,它讨论变量之间的关联。诸如“变量X与变量Y相关吗?”这个层次上的回答。然而,在第一级中还没有提到因果关系。第一级的例子是观察到公鸡啼叫与日出有关,然而,这种推理不能描述因果关系,因为我们不能说日出是否导致公鸡鸣叫。

因果关系阶梯的第二级被称为“干预”。在这一层面上的推理需要回答“如果我干预X,这将如何影响结果Y的概率?”例如,“吸烟会增加我患肺癌的几率吗?”存在于因果阶梯的第二级。这种推理调用了因果关系,可以用于分析比第一级推理更多的问题。

因果关系阶梯的第三级被称为“反事实”,包括回答一些问题:如果环境不同,会发生什么?这种推理在更大程度上调用了因果关系。书中给出的一个反事实问题是“如果奥斯瓦尔德没有杀死肯尼迪,他还会活着吗?”

第二章:从海盗到豚鼠:因果推断的起源

Chapter 2 starts with a brief summary of the contributions of Francis Galton and Karl Pearson to the development of statistics in the late 19th Century and early 20th Centuries. The authors blame Galton for keeping the study of statistics on the first rung of the ladder of causation and discouraging any discussion of causality in statistics. Causal analysis using path diagrams is then introduced through the explanations of the work of Sewall Wright.

第二章首先简要总结了弗朗西斯·高尔顿和卡尔·皮尔森在19世纪末20世纪初对统计学发展的贡献。作者们指责高尔顿把统计学的研究放在了因果关系阶梯的第一级,并阻止了统计学中任何关于因果关系的讨论。使用路径图的因果分析,然后介绍了通过解释Sewall Wright的工作。

Chapter 3: From Evidence to Causes: Reverend [Thomas Bayes|Bayes] meets Mr Holmes[edit]

Chapter 3 provides an introduction to Bayes Theorem. Then Bayesian Networks are introduced. Finally, the links between Baysian networks and causal diagrams are discussed.

Chapter 4: Confounding and Deconfounding, or, Slaying the Lurking Variable[edit]

This chapter introduces the idea of confounding and describes how causal diagrams can be used to identify confounding variables and determine their effect. Pearl explains that randomized controlled trials (RCTs) can be used to nullify the effect of confounders, but shows that, provided one has a causal model of confounding, an RCT does not necessarily have to be performed to get results.

Chapter 5: The Smoke-filled Debate: Clearing the Air[edit]

This chapter takes a historical approach to the question 'does smoking cause lung cancer?', focusing on the arguments made by Abraham Lilienfeld, Jacob Yerushalmy, Ronald Fisher and Jerome Cornfield. The authors explain that, though cigarette smoking was clearly correlated with lung cancer, some, such as Fisher and Yerushalmy, believed that the two variables were confounded and argued against the hypothesis that cigarettes caused the cancer. The authors then explain how causal reasoning (as developed in the rest of the book) can be used to argue that cigarettes do indeed cause cancer.

Chapter 6: Paradoxes Galore![edit]

This chapter examines several paradoxes, including the Monty Hall Problem, Simpson's paradox, Berkson's paradox and Lord's paradox. The authors show how these paradoxes can be resolved using causal reasoning.

Chapter 7: Beyond Adjustment: The Conquest of Mount Intervention[edit]

This chapter looks at the 'second rung' of the ladder of causation introduced in chapter 1. The authors describe how to use causal diagrams to ascertain the causal effect of performing interventions (eg. smoking) on outcomes (such as lung cancer). The 'front-door criterion' and the 'do-calculus' are introduced as tools for doing this. The chapter finishes with two examples, used to introduce the use of instrumental variables to estimate causal relationships. The first is John Snow's discovery that cholera is caused by unsanitary water supplies. The second is the relationship between cholesterol levels and likelihood of a heart attack.

Chapter 8: Counterfactuals: Mining worlds that could have been[edit]

This chapter examines the third rung of the ladder of causation: counterfactuals. The chapter introduces 'structural causal models', which allow reasoning about counterfactuals in a way that traditional (non-causal) statistics does not. Then, the applications of counterfactual reasoning are explored in the areas of climate science and the law.

Chapter 9: Mediation: The Search for Mechanism[edit]

This chapter discusses mediation: the mechanism by which a cause leads to an effect. The authors discuss the work of Barbara Stoddard Burks on the causes of intelligence of children, the 'algebra for all' policy by Chicago public schools, and the use of tourniquets to treat combat wounds.

Chapter 10: Big Data, Artificial Intelligence and the Big Questions[edit]

The final chapter discusses the use of causal reasoning in big data and artificial intelligence (AI) and the philosophical problem that AI would have to reflect on its own actions, which requires counterfactual (and therefore causal) reasoning.

致谢

Scientific Background, excerpts, errata, and a list of 37 reviews of The Book of Why is provided on Judea Pearl's web page

The Book of Why was reviewed by Jonathan Knee in The New York Times. The review was positive, with Knee calling the book "illuminating". However, he describes some parts of the book as "challenging", stating that the book is "not always fully accessible to readers who do not share the author's fondness for equations".

Tim Maudlin gave the book a mixed review in The Boston Review, calling it a "splendid overview of the state of the art in causal analysis". However, Maudlin criticizes the inclusion of "counterfactuals" as separate rung on the "ladder of causation", stating counterfactuals are so closely entwined with causal claims that it is not possible to think causally but not counterfactually". Maudlin also criticizes the section on free will for its "imprecision and lack of familiarity with the philosophical literature". Finally he points to the work of several scientists (including Clark Glymour) who developed similar ideas to Pearl, and claims that Pearl "could have saved himself literally years of effort had he been apprised of this work".

In a rebuttal, Pearl notes that, not only was he apprised of these scientists' work, but he actively collaborated in its creation. Additionally, the key developments described in The Book of Why, among them (1) identification analysis, (2) the algorithmization of counterfactuals, (3) mediation analysis, and (4) external validity, far surpass the narrow philosophical literature of the pre-2000 era.

Zoe Hackett, writing in Chemistry World, gave The Book of Why a positive review, with the caveat that "[i]t requires concentration, and a studious effort to work through the mind-bending statistical problems posited in the text". The review concludes by stating that "[t]his book is a must for any serious student of philosophy of science, and should be required reading for any first-year undergraduate statistics class".

Lisa R. Goldberg wrote a detailed, technical review in Notices of the American Mathematical Society.

参考文献

  1. Judea Pearl's information page for The Book of Why, http://bayes.cs.ucla.edu/WHY/
  2. "Review: The Book of Why Examines the Science of Cause and Effect". The New York Times. 1 June 2018.
  3. Tim Maudlin (4 September 2019). "The Why of the World". The Boston Review.
  4. Zoe Hackett (18 January 2019). "The Book of Why: The New Science of Cause and Effect". Chemistry World.
  5. Lisa R. Goldberg (August 2019). "The Book of Why" (PDF). Notices of the American Mathematical Society.