Research talk: Causal ML and fairness
Observing heterogeneous treatment effects across different demographic groups is an important mechanism for evaluating fairness. However, relatively little data is available for certain demographics, in which case researchers may combine multiple data sources to increase statistical power. The stakes are especially high in healthcare—it is imperative to accurately measure the effectiveness of treatments for diseases that could disproportionately impact underrepresented patient subgroups. Join researcher Allison Koenecke, from the Machine Learning & Statistics Group at Microsoft Research New England, to discuss federated causal inference. Because legal and privacy considerations may restrict individual-level information sharing across data sets, we introduce federated methods for treatment effect estimation that only utilize summary-level statistics from each data set. These asymptotically guaranteed methods provide variance estimates and doubly robust treatment effects under model assumptions on heterogeneous data sets.
Learn more about the 2021 Microsoft Research Summit: https://Aka.ms/researchsummit (opens in new tab)
- 轨迹:
- Causal Machine Learning
- 日期：
- 演讲者：
- Allison Koenecke
- 所属机构：
- Microsoft Research
-
-
Allison Koenecke
Postdoctoral Researcher
-
-
Causal Machine Learning
-
Opening remarks: Causal Machine Learning
Speakers:- Cheng Zhang
-
-
Research talk: Causal ML and business
Speakers:- Jacob LaRiviere
-
-
-
Panel: Challenges and opportunities of causality
Speakers:- Susan Athey,
- Yoshua Bengio,
- Judea Pearl
-
-
Research talk: Causal ML and fairness
Speakers:- Allison Koenecke
-
Panel: Causal ML Research at Microsoft
Speakers:- Daniel McDuff,
- Javier González,
- Justin Ding
-
Research talk: Post-contextual-bandit inference
Speakers:- Nathan Kallus
-
-
-
-
Panel: Causal ML in industry
Speakers:- Ya Xu,
- Totte Harinen,
- Dawen Liang
-
