University of Twente Student Theses


Domain adaptation under structural causal models

Kanger, Thomas (2023) Domain adaptation under structural causal models.

[img] PDF
Abstract:The statistical machine learning problem of domain adaptation (DA) is inspired by the human ability to transfer knowledge from one task to a different but similar task. DA arises when the training and test data are different. By exploiting the relationship between the training and test data, DA methods aim to predict the labels of test data. The performance of DA methods depends on the causal character between the covariates and the labels of training and test data. This paper quantified the expected error of the CIP estimator in the case of a standard linear causal model. Next, some experiments were done to quantify the model’s accuracy in labelling test data, while making some adjustments to keep the computation time low.
Item Type:Essay (Bachelor)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Programme:Applied Mathematics BSc (56965)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page