Causes, measurement, and mitigation of loss discrepancy
Abstract/Contents
- Abstract
- Machine learning models influence people's lives profoundly. In spite of their great performance, it has been observed that their predictions are often discriminatory against protected groups (e.g., women). One way to analyze the discrimination in machine learning models is to measure the difference in model performance across groups or individuals known as loss discrepancy. There are two flavors of loss discrepancies: statistical and counterfactual loss discrepancy. Statistical loss discrepancy measures how much-protected groups are impacted differently. Counterfactual loss discrepancy measures how much similar individuals are treated differently because of their group membership. Recent studies usually attribute this loss discrepancy to an information deficiency for one group (e.g., one group has less data). In this thesis, I show that: 1) Even when there is no information deficiency specific to one group (e.g., both groups have infinite data), adding the same amount of feature noise to all individuals leads to loss discrepancy; 2) Even when features are noiseless and perfectly determine the prediction target, the inductive bias in the overparameterized regime leads to reliance on spurious features and loss discrepancy. Understanding the source of loss discrepancies helps to come up with mitigation methods. I explain two methods to mitigate loss discrepancy. First, I explain how to leverage unlabeled data to reduce counterfactual loss discrepancy without affecting accuracy. Next, I propose another mitigation method, the unanimity principle: only predict when all models consistent with the training data predict the same output. I operationalize this principle for semantic parsing, the task of mapping utterances to logical forms. I develop a simple, efficient method that reasons over the infinite set of consistent models by only checking two models. I prove that this method obtains 100% precision, thus no loss discrepancy. Finally, I investigate the methods to measure of loss discrepancy when there is no information about protected attributes, or there are exponentially many protected groups. I introduce and study a notion which I call maximum weighted loss discrepancy (MWLD), the maximum (weighted) difference between the loss of a group and the loss of the population. I show that it is statistically impossible to estimate MWLD when all groups have equal weights, but for a particular family of weighting functions, I show how to estimate MWLD efficiently. Finally, I draw a relation between MWLD and loss variance, a quantity that arises in generalization bounds.
Description
Type of resource | text |
---|---|
Form | electronic resource; remote; computer; online resource |
Extent | 1 online resource. |
Place | California |
Place | [Stanford, California] |
Publisher | [Stanford University] |
Copyright date | 2021; ©2021 |
Publication date | 2021; 2021 |
Issuance | monographic |
Language | English |
Creators/Contributors
Author | Khani, Fereshte |
---|---|
Degree supervisor | Liang, Percy |
Thesis advisor | Liang, Percy |
Thesis advisor | Hashimoto, Tatsunori |
Thesis advisor | Reingold, Omer |
Degree committee member | Hashimoto, Tatsunori |
Degree committee member | Reingold, Omer |
Associated with | Stanford University, Computer Science Department |
Subjects
Genre | Theses |
---|---|
Genre | Text |
Bibliographic information
Statement of responsibility | Fereshte Khani. |
---|---|
Note | Submitted to the Computer Science Department. |
Thesis | Thesis Ph.D. Stanford University 2021. |
Location | https://purl.stanford.edu/gw991vt5365 |
Access conditions
- Copyright
- © 2021 by Fereshte Khani
Also listed in
Loading usage metrics...