Abstract
Experts advising decision-makers are likely to display expertise which varies as a function of the problem instance. In practice, this may lead to sub-optimal or discriminatory decisions against minority cases. In this work, we model such changes in depth and breadth of knowledge as a partitioning of the problem space into regions of differing expertise. We provide here new algorithms that explicitly consider and adapt to the relationship between problem instances and experts’ knowledge. We first propose and highlight the drawbacks of a naive approach based on nearest neighbor queries. To address these drawbacks we then introduce a novel algorithm — expertise trees — that constructs decision trees enabling the learner to select appropriate models. We provide theoretical insights and empirically validate the improved performance of our novel approach on a range of problems for which existing methods proved to be inadequate.
Original language | English |
---|---|
Title of host publication | Proceedings of the 40th International Conference on Machine Learning |
Publisher | JMLR |
Pages | 79-90 |
Number of pages | 12 |
Volume | 202 |
ISBN (Print) | 2640-3498 |
Publication status | Published - Jul 2023 |
Event | International Conference on Machine Learning - Hawaiʻi Convention Center, Honolulu, United States Duration: 23 Jul 2023 → 29 Jul 2023 https://icml.cc |
Publication series
Name | Proceedings of Machine Learning Research |
---|
Conference
Conference | International Conference on Machine Learning |
---|---|
Country/Territory | United States |
City | Honolulu |
Period | 23/07/23 → 29/07/23 |
Internet address |
Bibliographical note
Funding Information:A.A. is supported by a FRIA grant (nr. 5200122F) by the National Fund for Scientific Research (F.N.R.S.) of Belgium. T.L. is supported by the F.N.R.S. project with grant numbers 31257234 and 40007793, the F.W.O. project with grant nr. G.0391.13N, the Service Public de Wallonie Recherche under grant n°2010235–ARIAC by DigitalWallonia4.ai. T.L and A.N. benefit from the support of the Flemish Government through the AI Research Program. T.L., V.T. and A.N. acknowledge the support by TAILOR, a project funded by EU Horizon 2020 research and innovation program under GA No 952215. V.T. acknowledges the support by HACID, a project funded by EU Horizon Europe research and innovation program under GA No 101070588. The resources and services used in this work were provided by the VSC (Flemish Supercomputer Center), funded by the Research Foundation - Flanders (FWO) and the Flemish Government.
Publisher Copyright:
© 2023 Proceedings of Machine Learning Research. All rights reserved.