Learning with privileged and sensitive information: a gradient-boosting approach
We consider the problem of learning with sensitive features under the privileged information setting where the goal is to learn a classifier that uses features not available (or too sensitive to collect) at test/deployment time to learn a better model at training time. We focus on tree-based learner...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2023-11-01
|
| Series: | Frontiers in Artificial Intelligence |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/frai.2023.1260583/full |
| _version_ | 1797628983580295168 |
|---|---|
| author | Siwen Yan Phillip Odom Rahul Pasunuri Kristian Kersting Sriraam Natarajan |
| author_facet | Siwen Yan Phillip Odom Rahul Pasunuri Kristian Kersting Sriraam Natarajan |
| author_sort | Siwen Yan |
| collection | DOAJ |
| description | We consider the problem of learning with sensitive features under the privileged information setting where the goal is to learn a classifier that uses features not available (or too sensitive to collect) at test/deployment time to learn a better model at training time. We focus on tree-based learners, specifically gradient-boosted decision trees for learning with privileged information. Our methods use privileged features as knowledge to guide the algorithm when learning from fully observed (usable) features. We derive the theory, empirically validate the effectiveness of our algorithms, and verify them on standard fairness metrics. |
| first_indexed | 2024-03-11T10:46:53Z |
| format | Article |
| id | doaj.art-5221005c5eba4d62893651d211111a2f |
| institution | Directory Open Access Journal |
| issn | 2624-8212 |
| language | English |
| last_indexed | 2024-03-11T10:46:53Z |
| publishDate | 2023-11-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Artificial Intelligence |
| spelling | doaj.art-5221005c5eba4d62893651d211111a2f2023-11-14T02:21:05ZengFrontiers Media S.A.Frontiers in Artificial Intelligence2624-82122023-11-01610.3389/frai.2023.12605831260583Learning with privileged and sensitive information: a gradient-boosting approachSiwen Yan0Phillip Odom1Rahul Pasunuri2Kristian Kersting3Sriraam Natarajan4Computer Science Department, University of Texas at Dallas, Dallas, TX, United StatesGeorgia Tech Research Institute, Georgia Institute of Technology, Atlanta, GA, United StatesAmazon, Seattle, WA, United StatesDepartment of Computer Science, Hessian Center for AI (hessian.AI), Technical University of Darmstadt, Darmstadt, GermanyComputer Science Department, University of Texas at Dallas, Dallas, TX, United StatesWe consider the problem of learning with sensitive features under the privileged information setting where the goal is to learn a classifier that uses features not available (or too sensitive to collect) at test/deployment time to learn a better model at training time. We focus on tree-based learners, specifically gradient-boosted decision trees for learning with privileged information. Our methods use privileged features as knowledge to guide the algorithm when learning from fully observed (usable) features. We derive the theory, empirically validate the effectiveness of our algorithms, and verify them on standard fairness metrics.https://www.frontiersin.org/articles/10.3389/frai.2023.1260583/fullprivileged informationfairnessgradient boostingknowledge-based learningsensitive features |
| spellingShingle | Siwen Yan Phillip Odom Rahul Pasunuri Kristian Kersting Sriraam Natarajan Learning with privileged and sensitive information: a gradient-boosting approach Frontiers in Artificial Intelligence privileged information fairness gradient boosting knowledge-based learning sensitive features |
| title | Learning with privileged and sensitive information: a gradient-boosting approach |
| title_full | Learning with privileged and sensitive information: a gradient-boosting approach |
| title_fullStr | Learning with privileged and sensitive information: a gradient-boosting approach |
| title_full_unstemmed | Learning with privileged and sensitive information: a gradient-boosting approach |
| title_short | Learning with privileged and sensitive information: a gradient-boosting approach |
| title_sort | learning with privileged and sensitive information a gradient boosting approach |
| topic | privileged information fairness gradient boosting knowledge-based learning sensitive features |
| url | https://www.frontiersin.org/articles/10.3389/frai.2023.1260583/full |
| work_keys_str_mv | AT siwenyan learningwithprivilegedandsensitiveinformationagradientboostingapproach AT phillipodom learningwithprivilegedandsensitiveinformationagradientboostingapproach AT rahulpasunuri learningwithprivilegedandsensitiveinformationagradientboostingapproach AT kristiankersting learningwithprivilegedandsensitiveinformationagradientboostingapproach AT sriraamnatarajan learningwithprivilegedandsensitiveinformationagradientboostingapproach |