We argue for supplementing the process of training a prediction algorithm by setting up a scheme for detecting the moment when the distribution of the data changes and the algorithm needs to be retrained. Our proposed schemes are based on exchangeability martingales, i.e., processes that are martingales under any exchangeable distribution for the data. Our method, based on conformal prediction, is general and can be applied on top of any modern prediction algorithm. Its validity is guaranteed, and in this paper we make first steps in exploring its efficiency.
|Title of host publication||Proceedings of Machine Learning Research|
|Editors||Lars Carlsson, Zhiyuan Luo, Giovanni Cherubin, Khuong Nguyen|
|Number of pages||20|
|Publication status||Published - 2021|
- conformal test martingales
- dataset shift
- IID assumption