Download - Causality Algorithms Virtual Reading Group
![Page 1: Causality Algorithms Virtual Reading Group](https://reader034.vdocuments.site/reader034/viewer/2022042318/625cd472dffee0779f7da913/html5/thumbnails/1.jpg)
Causality & Algorithms Virtual Reading Group
June 19, 2020
![Page 2: Causality Algorithms Virtual Reading Group](https://reader034.vdocuments.site/reader034/viewer/2022042318/625cd472dffee0779f7da913/html5/thumbnails/2.jpg)
Why are we meeting?
•Goal: understand current work in causal inference and figure out interesting questions from a TCS perspective.
•Make concrete connections to property testing? non‐asymptotic bounds/sample complexity? robust statistics? approximation algorithms? hardness? Also, (re‐)defining things in more CS‐friendly language.
![Page 3: Causality Algorithms Virtual Reading Group](https://reader034.vdocuments.site/reader034/viewer/2022042318/625cd472dffee0779f7da913/html5/thumbnails/3.jpg)
Administrivia
• Plan is to meet once every two weeks.
• Please volunteer! You don’t necessarily have to be an expert on the topic. The goal is to learn and discuss.
• I will post video recordings of the meetings.
![Page 4: Causality Algorithms Virtual Reading Group](https://reader034.vdocuments.site/reader034/viewer/2022042318/625cd472dffee0779f7da913/html5/thumbnails/4.jpg)
Individual Treatment Effect Estimation & Causal Forests
Causality & Algorithms Virtual Reading GroupArnab Bhattacharyya
June 19, 2020
![Page 5: Causality Algorithms Virtual Reading Group](https://reader034.vdocuments.site/reader034/viewer/2022042318/625cd472dffee0779f7da913/html5/thumbnails/5.jpg)
Meta‐learners for Estimating Heterogeneous Treatment Effects using Machine Learning. Kunzel, Sekhon, Bickel, Yu. PNAS, 116 (10), pg. 4156—4165, 2019.
𝑚 𝑥
𝑚 𝑥
![Page 6: Causality Algorithms Virtual Reading Group](https://reader034.vdocuments.site/reader034/viewer/2022042318/625cd472dffee0779f7da913/html5/thumbnails/6.jpg)
𝐼𝑇𝐸 𝑥
𝐼𝑇𝐸 𝑥
Meta‐learners for Estimating Heterogeneous Treatment Effects using Machine Learning. Kunzel, Sekhon, Bickel, Yu. PNAS, 116 (10), pg. 4156—4165, 2019.
X‐learner
![Page 7: Causality Algorithms Virtual Reading Group](https://reader034.vdocuments.site/reader034/viewer/2022042318/625cd472dffee0779f7da913/html5/thumbnails/7.jpg)
Estimation and Inference of Heterogeneous Treatment Effects using Random Forests. Wager & Athey. Journal of the American Statistical Association, 113:523, pg. 1228—1242, 2018.
𝐼𝑇𝐸 𝑥1
| 𝑖:𝑇 1,𝑋 ∈ 𝐿 𝑥 | 𝑌∈ : ,∈
1| 𝑖:𝑇 0,𝑋 ∈ 𝐿 𝑥 | 𝑌
∈ : ,∈
Choose split so as to maximize:
𝐼𝑇𝐸 𝑋∈
![Page 8: Causality Algorithms Virtual Reading Group](https://reader034.vdocuments.site/reader034/viewer/2022042318/625cd472dffee0779f7da913/html5/thumbnails/8.jpg)
Estimation and Inference of Heterogeneous Treatment Effects using Random Forests. Wager & Athey. Journal of the American Statistical Association, 113:523, pg. 1228—1242, 2018.