| |
|
Description: |
In the application of machine learning to real life decision-making systems, e.g., credit scoring and criminal justice, the prediction outcomes might discriminate against people with sensitive attributes, leading to unfairness. The commonly used strategy in fair machine learning is to include fairness as a constraint or a penalization term in the minimization of the prediction loss, which ultimately limits the information given to decision-makers. In this talk, we introduce a new approach to handle fairness by formulating stochastic multi-objective optimization (SMOO) problems for which the corresponding Pareto fronts uniquely and comprehensively define the accuracy-fairness trade-offs.
We will the talk start by giving a brief description of the main goals in Data Analysis and Learning, with an emphasis on the type of optimization problems needed to be solved. Then, we will review the classical stochastic gradient method for (single-objective) stochastic optimization. SMOO appears when optimizing conflicting functions which involve data that is uncertain or unknown. We will also review the stochastic multi-gradient method, seen as a natural extension of the classical stochastic gradient method from single to multi-objective optimization.
|
|
Date: |
|
| Start Time: |
14:30 |
|
Speaker: |
Luís Nunes Vicente (Lehigh Univ., USA & CMUC, Univ. Coimbra)
|
|
Institution: |
Lehigh University and CMUC
|
|
Place: |
Remote seminar via Zoom https://videoconf-colibri.zoom.us/j/88350027907
|
| Research Groups: |
-Numerical Analysis and Optimization
|
|
See more:
|
<Main>
|
|
