Ececal at 13:03, 20 September 2022

2022-09-20T13:03:49Z

Ececal: Our goal is to design algorithms using conformal prediction framework that make fair predictions across various groups based on e.g., age, sex, income.

2022-09-20T12:41:43Z

Our goal is to design algorithms using conformal prediction framework that make fair predictions across various groups based on e.g., age, sex, income.

New page

{{StudentProjectTemplate
|Summary=Our goal is to design algorithms using conformal prediction framework that make fair predictions across various groups based on e.g., age, sex, income.
|Keywords=fair machine learning, bias, fairness
|TimeFrame=Fall 2022
|References=https://ojs.aaai.org/index.php/AAAI/article/view/21459
|Supervisor=Ece Calikus,
|Level=Master
|Status=Open
}}
The goal of fairness in machine learning is to design algorithms that make fair predictions across various demographic groups. Conformal prediction is a technique devised to assess the uncertainty of predictions produced by a machine learning model. In particular, given an input, conformal prediction estimates a prediction interval in regression problems and a set of classes in classification problems. Both the prediction interval and sets are guaranteed to cover the true value with high probability.

Our goal is to use this approach to assess prediction uncertainty among different subpopulations (i.e., bias) with sensitive attributes (e.g., race, sex, income, age, etc.) and extend this approach to provide equal coverage among different subgroupsPotential applications for this project involve healthcare, education, and the environment.

@@ Line 4: / Line 4: @@
 |TimeFrame=Fall 2022
 |References=https://ojs.aaai.org/index.php/AAAI/article/view/21459
 |Supervisor=Ece Calikus,
 |Level=Master
 |Status=Open
 }}
-The goal of fairness in machine learning is to design algorithms that make fair predictions across various demographic groups. Conformal prediction is a technique devised to assess the uncertainty of predictions produced by a machine learning model. In particular, given an input, conformal prediction estimates a prediction interval in regression problems and a set of classes in classification problems. Both the prediction interval and sets are guaranteed to cover the true value with high probability.
+The goal of fairness in machine learning is to design algorithms that make fair predictions across various demographic groups. Conformal prediction [1] is a technique devised to assess the uncertainty of predictions produced by a machine learning model. In particular, given an input, conformal prediction estimates a prediction interval in regression problems and a set of classes in classification problems. Both the prediction interval and sets are guaranteed to cover the true value with high probability.
 Our goal is to use this approach to assess prediction uncertainty among different subpopulations (i.e., bias) with sensitive attributes (e.g., race, sex, income, age, etc.) and extend this approach to provide equal coverage among different subgroupsPotential applications for this project involve healthcare, education, and the environment.

Fair Conformal Prediction - Revision history

Ececal at 13:03, 20 September 2022

Ececal: Our goal is to design algorithms using conformal prediction framework that make fair predictions across various groups based on e.g., age, sex, income.